university of alberta library release form

124
University of Alberta Library Release Form Name of Author: Marek Welz Title of Thesis: Blast Monitoring at Diavik Diamond Mines Inc. to Protect Fish Habitat Degree: Master of Science Year this Degree Granted: 2006 Permission is hereby granted to the University of Alberta Library to reproduce single copies of this thesis and to lend or sell such copies for private, scholarly or scientific research purposes only. The author reserves all other publication and other rights in association with the copyright in the thesis, and except as herein before provided, neither the thesis nor any substantial portion thereof may be printed or otherwise reproduced in any material form whatever without the author’s prior written permission. ________________________ Marek Welz 228-D CEB University of Alberta, Edmonton, Alberta Canada T6G 2G7

Upload: others

Post on 25-Apr-2022

1 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: University of Alberta Library Release Form

University of Alberta

Library Release Form

Name of Author: Marek Welz Title of Thesis: Blast Monitoring at Diavik Diamond Mines Inc. to Protect Fish Habitat Degree: Master of Science Year this Degree Granted: 2006 Permission is hereby granted to the University of Alberta Library to reproduce single copies of this thesis and to lend or sell such copies for private, scholarly or scientific research purposes only. The author reserves all other publication and other rights in association with the copyright in the thesis, and except as herein before provided, neither the thesis nor any substantial portion thereof may be printed or otherwise reproduced in any material form whatever without the author’s prior written permission. ________________________ Marek Welz 228-D CEB University of Alberta, Edmonton, Alberta Canada T6G 2G7

Page 2: University of Alberta Library Release Form

University of Alberta

Blast Monitoring at Diavik Diamond Mines Inc. to Protect Fish Habitat

by

Marek Welz

A thesis submitted to the Faculty of Graduate Studies and Research in partial

fulfillment of the requirements for the degree of Master of Science

in

Geophysics

Department of Physics

Edmonton, Alberta

Fall 2006

Page 3: University of Alberta Library Release Form

University of Alberta

Faculty of Graduate Studies and Research

The undersigned certified that they have read, and recommend to the Faculty of Graduate Studies and Research for acceptance, a thesis entitled Blast Monitoring at Diavik Diamond Mines Inc. to Protect Fish Habitat submitted by Marek Welz in partial fulfillment of the requirements for the degree of Master of Science.

______________________________ Dr. D. Schmitt (Supervisor)

______________________________ Dr. T.G. Joseph

______________________________ Dr. V. A. Kravchinsky

______________________________ Dr. W.M. Tonn

Date:

Page 4: University of Alberta Library Release Form

Dedication

To my mother.

Page 5: University of Alberta Library Release Form

Acknowledgements

I am indebted to Doug Schmitt; for his supervision, guidance and his hard work.

Without him I would not have been able to complete this work. The work was

collaboration with Sean Faulkner and Dr. Bill Tonn of the Department of Biological

Sciences. I thank them for their expertise on fishes and for the support and many

discussions about the project. This research was made possible through a research grant

and in-kind support provided by Diavik Diamond Mines Inc. (DDMI). I would also like

to thank the geophysics group and the support within it; Especially Len Tober for his

technical expertise, and Dan Collis for his help in the initial deployment.

Vista 5.1 seismic processing software provided by Gedco of Calgary was used in

processing of some of the data.

Page 6: University of Alberta Library Release Form

Abstract:

Diavik Diamond Mines Inc. is conducting open pit and underground mining of

kimberlites within close proximity of a large freshwater lake that supports populations of

fishes and other aquatic organisms. As the principal method of excavation is through

blasting; the resulting blast wave is expected to expose incubating fish eggs to peak

particle velocities (PPVs) greater than guidelines suggested by the Department of

Fisheries and Oceans (DFO). Blasts were monitored using specialised 4 – component

underwater sensors which measured the particle velocities and overpressures

simultaneously. PPVs were related to mortality in fish eggs both in the lake and in a

laboratory setting where blasting events were simulated. Standard blast monitoring

techniques were used to analyse data and create scaled distance relations in order to

predict the attenuation of PPVs. Results show that attenuation was greater than initially

expected. Further analysis showed no changes in blasting effects as the pit deepened.

Page 7: University of Alberta Library Release Form

Table of Contents

Chapter 1) Introduction:...................................................................................................... 1

1.1) Objectives: ............................................................................................................... 3

Chapter 2) Background....................................................................................................... 5

2.1) Basic Physical Principles of Seismic Waves........................................................... 6

2.1.1) Basic Wave Propagation................................................................................... 7

2.1.2) Elastic Moduli................................................................................................... 9

2.1.3) Wave Energy .................................................................................................. 17

2.1.4) Decay of Amplitude........................................................................................ 19

i) Geometrical Spreading .......................................................................................... 19

ii) Intrinsic attenuation.............................................................................................. 21

iii) Attenuation at geologic boundaries..................................................................... 22

iv) Scattering Attenuation ......................................................................................... 23

v) Combination of effects ......................................................................................... 24

2.2) Introduction to Blast Monitoring........................................................................... 27

2.2.1) Peak Particle Velocity .................................................................................... 30

2.2.2) Overpressure................................................................................................... 32

2.2.3) Scaled Distance............................................................................................... 33

2.3) Summary................................................................................................................ 36

Chapter 3) Diavik site overview ....................................................................................... 37

3.1) Geological Setting ................................................................................................. 38

3.2) Study site ............................................................................................................... 41

Page 8: University of Alberta Library Release Form

3.3) Geophysical site characterisation .......................................................................... 41

3.3.4) Refraction seismic survey............................................................................... 42

3.3.5) Passive seismic survey.................................................................................... 44

3.4) Predicted Blast Zone.............................................................................................. 45

3.4.1) Peak Particle Velocity .................................................................................... 46

3.4.2) Over Pressure.................................................................................................. 48

3.5) Summary................................................................................................................ 50

Chapter 4) Methodology; Field component...................................................................... 52

4.1) Site Selection ......................................................................................................... 53

4.2) Sensor Placement................................................................................................... 55

4.3) Equipment.............................................................................................................. 56

4.4) Data Collection...................................................................................................... 57

4.5) Data Processing and analysis................................................................................. 61

Chapter 5) Field Component Results................................................................................ 62

5.1) Analysis ................................................................................................................. 62

5.2) DDMI Prediction Function.................................................................................... 63

5.3) South Dike (8902) ................................................................................................. 65

5.4) Reference Island (8903)......................................................................................... 66

5.5) Tern Island (8904) ................................................................................................. 66

5.6) East Dike (8905).................................................................................................... 66

5.7) Blast Zone.............................................................................................................. 70

5.8) Changes in effects as the pit deepens .................................................................... 72

5.8.1) Changes in normalised PPV ........................................................................... 74

Page 9: University of Alberta Library Release Form

5.8.2) Changes in the Scaled Distance function ....................................................... 76

Chapter 6) Laboratory Study ............................................................................................ 79

6.1) Methodology.......................................................................................................... 80

6.2) Results ................................................................................................................... 81

6.3) Pseudo energy........................................................................................................ 83

Chapter 7) Discussion....................................................................................................... 86

7.1) Field component .................................................................................................... 86

7.2) Changes in the blasting effects .............................................................................. 90

Chapter 8) Conclusions..................................................................................................... 92

References......................................................................................................................... 96

Appendices........................................................................................................................ 99

Appendix 1: Response curves for the Underwater sensor ............................................ 99

Appendix 2: Blast Mate blasting record ..................................................................... 101

Appendix 3: Technical Specifications for the Instantel Minimate Plus Series III……...102

Appendix 4: Blasting Data.......................................................................................... 103

Page 10: University of Alberta Library Release Form

List of Tables

Table 1) Current Alberta seismic exploration regulation offsets, Schedule 2 of Alberta

Government (2005)........................................................................................................... 29

Page 11: University of Alberta Library Release Form

List of Figures

Figure 1-1) Air photo showing the proposed pit (A154) location within the drained area

near East Island on Lac de Gras. The A154N and A154S kimberlitic pipes are indicated

in red. The inset map shows the location of DDMI in Canada’s Northwest Territories.

Courtesy, DDMI. ................................................................................................................ 2

Figure 2-1) Poisson ratio plotted against the Vp/Vs ratio................................................. 13

Figure 2-2) The Vr/Vs ratio plotted against the Vp/Vs ratio. ........................................... 14

Figure 2-3) Retrograde particle motion of a Rayleigh wave. Wave propagation direction

along the surface is represented by the large horizontal arrow. At a given point the

particles move ‘backwards’ around the ellipse as shown. ................................................ 15

Figure 2-4) Illustration of the relative decay of a cylindrically spreading surface wave and

a spherically spreading body wave with distance relative to the (arbitrary) reference

distance of 0.1 m for a wave speed of 5000 m/s and a frequency of 100 Hz. Relative

amplitudes report in dB10. Numerical values represent quality factors included in the

calculations. Q = ∞ means no attenuation. ...................................................................... 26

Figure 2-5) Illustration of the definition of the PPV on a particle velocity versus time plot

where the peak particle velocity is the maximum excursion from zero amplitude. ......... 28

Figure 3-1) Simplified geology at Diavik. The blue background is Lac de Gras, DDMI’s

four kimberlie pipes (A154N, A154S, A418 and A21) are all situated within the waters of

the lake very close to East Island where the operations are situated. (Modified from

Stubley Geoscience, 1998)................................................................................................ 39

Figure 3-2) Figure showing the location of the seismic line receiver array and the source.

........................................................................................................................................... 42

Page 12: University of Alberta Library Release Form

Figure 3-3) Frequency spectrum of refraction seismic record. The record contains,

seismically speaking, very high frequencies..................................................................... 43

Figure 3-4) A shot from the refraction survey. Surface wave velocity is in the range of

1000 m/s. Data is displayed after ormsby filtering (10, 15, 600, 800) and RMS scaling. 43

Figure 3-5) The first arrival from the blasting event. A best fit line defines the velocity to

be around 2600 m/s........................................................................................................... 44

Figure 3-6) Frequency spectrum for the passively recorded blasting event. .................... 45

Figure 3-7) Map of area showing proposed fish incubator sites (black squares). Circular

contours represent distance from the centre of A154N/S, contours each 500-m and the red

line indicates the DDMI predicted zone where PPVs would exceed DFO guideline levels.

........................................................................................................................................... 47

Figure 4-1) Locations of the 4 sensors in relation to the dike, proposed pit and predicted

Blast Zone (Red Line). Distance between scale marks is 1000 m of the UTM grid. DDMI

sensors are deployed at various locations along the dike. Note that the Blast Zone takes

into account the location of the next pit south and east of the original. ........................... 54

Figure 4-2) Placement of the 4-C sensor. The sensor is attached to the chain and also

placed under rocks to ensure coupling.............................................................................. 56

Figure 4-3) A cutaway view of the OYO 4 Component sensor. The gimballed geophones

can be seen and the hydrophone is situated at the left end of the sensor.......................... 57

Figure 4-4) Typical waveform due to a blast as presented using the Instantel Blast

WareTM software. The channels are overpressure (top), and the 3 geophone channels,

longitudinal, vertical and transverse. The vertical scale is 10 mm/s for the PPV and 10 pa

for the overpressure per division. Inset) Frequency spectrum for each of the geophone

Page 13: University of Alberta Library Release Form

channels; Y-axis Particle Velocity (mm/s) and, X-axis Frequency in Hz. Both Axes are

log scale. ........................................................................................................................... 59

Figure 4-5) Blast record for a non blasting related event. As above, the channels are

overpressure (top), and the 3 geophone channels, longitudinal, vertical and transverse.

The vertical scale is 2 mm/s for the PPV and 10 pa for the overpressure per division. ... 60

Figure 4-6) A typical waveform due to electrical noise. As above, the channels are

overpressure (top), and the 3 geophone channels, longitudinal, vertical and transverse.

The vertical scale is 10 mm/s for the PPV and 10 pa for the overpressure per division. . 60

Figure 5-1) The DDMI prediction function (Equation 5-2) presented in red lines with the

data collected by DDMI from August 27, 2002 to January 17, 2004. The black lines

represent the prediction function derived from the plotted data (Equation 5-2)............... 64

Figure 5-2) Scaled Distance plot of the four U of A sensors compared to the DDMI data.

The red lines represent the minimum, median and maximum PPV for a given scaled

distance from the DDMI prediction function. 8902, 8903, 8904 & 8905 are South Dike,

Reference Island, Tern Island and East Dike respectively. .............................................. 64

Figure 5-3) Logarithmic plot of the measured particle velocity and scaled distance for the

South Dike location. The lines indicate the minimum, median and maximum PPVs to be

expected at the site (r-square .26). .................................................................................... 67

Figure 5-4) Logarithmic plot of the measured particle velocity and scaled distance for the

Reference Island location. The lines indicate the minimum, median and maximum PPVs

to be expected at the site(r-square .09). ............................................................................ 67

Page 14: University of Alberta Library Release Form

Figure 5-5) Logarithmic plot of the measured particle velocity and scaled distance for the

Tern Island location. The lines indicate the minimum, median and maximum PPVs to be

expected at the site(r-square .27). ..................................................................................... 68

Figure 5-6) Logarithmic plot of the measured particle velocity and scaled distance for the

East Dike location. The lines indicate the minimum, median and maximum PPVs to be

expected at the site(r-square .26). ..................................................................................... 68

Figure 5-7) Minimum, maximum and median predicted PPVs plotted along with actual

recorded values for South Dike. The line indicates the DFO guideline value of 13 mm/s.

........................................................................................................................................... 69

Figure 5-8) Minimum, maximum and median predicted PPVs plotted along with actual

recorded values for Reference Island. Note that the vertical scale is much smaller in this

case.................................................................................................................................... 69

Figure 5-9) Minimum, maximum and median predicted PPVs plotted along with actual

recorded values for Tern Island. The line indicates the DFO guideline value of 13 mm/s.

........................................................................................................................................... 69

Figure 5-10) Minimum, maximum and median predicted PPVs plotted along with actual

recorded values for East Dike. The line indicates the DFO guideline value of 13 mm/s. 70

Figure 5-11) Logarithmic plot of the measured particle velocity and scaled distance for

the East Dike, South Dike and Reference Island combined. The lines indicate the

minimum, median and maximum PPVs to be expected (r-square .24). ........................... 71

Figure 5-12) PPV predictions for different charge weights and distances based on data

from the 3 near dike sites. The dashed line represents 13 mm/s. The Predicted values are

based on the upper bound of the function......................................................................... 72

Page 15: University of Alberta Library Release Form

Figure 5-13) Blast elevations at DDMI since January 2003. Over time the elevations have

decreased as the open pit operation increases in depth..................................................... 73

Figure 5-14) PPVs (mm/s) as measured by DDMI sensors since January 2003. ............. 74

Figure 5-15) Normalised (DDMI) PPVs plotted against the date..................................... 75

Figure 5-16) Normalised (DDMI) PPVs plotted against the blast elevation. ................... 75

Figure 5-17) Normalised (DDMI) PPVs plotted against distance from blast. As the

distance increases, the measured PPV decreases.............................................................. 75

Figure 5-18) Normalised (DDMI) PPVs plotted against the blast weight........................ 76

Figure 5-19) Change in A & M over time. (DDMI data). The scalar variable, A, is on the

left axis and M, the exponential variable is on the right axis. .......................................... 77

Figure 5-20) Change in A & M as the depth increases (elevation decreases). (DDMI

data). The scalar variable, A, is on the left axis and M, the exponential variable is on the

right axis............................................................................................................................ 77

Figure 5-21) Change in A & M as the distance from the blast increases. (DDMI data).

The scalar variable, A, is on the left axis and M, the exponential variable is on the right

axis. ................................................................................................................................... 77

Figure 6-1) The weight-drop apparatus; a: steel pole, secured at top with straps (not

shown); b: pin to release the weight apparatus; c: the weight set-up, including an

aluminium sleeve and cast iron weights; d: holes drilled in the pole spaced 15 cm apart; e:

0.64 cm thick rubber mat; f: 0.64 cm thick steel base plate; g: 5 cm foam padding; h: 2.5

cm rubber mat, and i: fibreglass tank................................................................................ 81

Figure 6-2) PPV = 138 mm/s Height = Standard Deviation (2.4) .................................... 82

Figure 6-3) PPV = 245 mm/s Height = Standard Deviation (3.0) .................................... 82

Page 16: University of Alberta Library Release Form

Figure 6-4) Blast records for a lab simulation event (a) and DDMI blasting event (b). The

bottom panel shows the 3 component wave form and the top panel shows the pseudo

energy calculated from the vector sum of the waveform. Note that the amplitude scales

are the same for comparison sake but the time scales are different. The DDMI event is

much longer than the lab event. ........................................................................................ 85

Figure 7-1) Comparison of scaled distance data over two periods shows that the

distribution has stayed the same. The solid line indicates 13 mm/s. ................................ 90

Page 17: University of Alberta Library Release Form

List of Variables

PPV Peak Particle Velocity, generally measured in mm/s

ω Angular Frequency (2πf)

A The displacement peak Amplitude

P Pressure in Pascals

VP/VS P & S wave velocities

ρ Material density

E Average over one wavelength of the total energy conveyed per unit wavefront

Z acoustic impedance (ρV)

PV Particle velocity

R / r Radius

y, y’, y’’ Particle displacement, velocity and acceleration

Q Quality factor

α Attenuation coefficient

SHi(t) Time dependent particle velocity for one component of the waveform

W Charge weight per delay

C Seismic wave velocity

SD Scaled distance (distance/(charge weight per delay)1/2

ML Richter Scale magnitude

t Time (usually in seconds)

M Scalar variable in the prediction function

A Exponential variable in the prediction function

Page 18: University of Alberta Library Release Form

1

Chapter 1) Introduction:

Regulations are in place to limit negative effects that can be experienced due energy

being released from blasting operations. In particular limits have been placed on the

maximum Peak Particle Velocity (PPV) levels to prevent negative cultural or

environmental effects. Prediction of maximum PPV is, therefore, important in

determining the maximum size of an explosion that can be used without damaging

anything at given distances.

Scaled distance relations that normalise for distance and blast weight per delay are

employed in predicting PPVs at given distances. These relations are empirical and must

be used carefully but can be very effective in PPV prediction once a reasonable data set is

collected. At Diavik Diamond Mines Inc. (DDMI) a scaled distance relation provided by

M. Davachi (Personal comm. DDMI, 1998a.) and calculations based on Department of

Fisheries and Oceans (DFO) guidelines both suggested that PPVs in excess of the DFO

mandated limit of 13mm/s would extend into Lac de Gras. Consequently, a Fisheries Act

section 32 Authorization (SC98001) relating to this activity requires that DDMI monitor

peak particle velocities and overpressures resulting from blasting activities in and around

the predicted blast zone, to define the actual blast zone; effects of blasting on fish must

also be determined. In particular, any incubating fish eggs within the blast zone could be

susceptible to mortality or damage, especially since they would be exposed to the

excessive velocities or pressures for 6-10 months. Because there is considerable

uncertainty regarding the cumulative effects of repeated guideline-exceeding forces on

over-wintering, developing eggs in Arctic lakes, and because the use of explosives in or

near arctic Canadian fisheries waters will likely increase in the future, it is important to

Page 19: University of Alberta Library Release Form

2

conduct a study into effects on fish and especially fish eggs, both with respect to the

A154 pit (Figure 1-1.) but also to assess more generally the relationship between the

current guidelines and effects under arctic field conditions.

Figure 1-1) Air photo showing the proposed pit (A154) location within the drained area near East Island on Lac de Gras. The A154N and A154S kimberlitic pipes are indicated in red. The inset map shows the location of DDMI in Canada’s Northwest Territories. Courtesy, DDMI.

This project is divided into two main components, a field study and a laboratory

study. The first component of the field study was monitoring and recording the effects of

blasting events on lake trout eggs that were placed beside the sensor. The second portion

of the field study, is to monitor as many blasting events as possible to assess the changes

in effects on the lake substrate as the pit deepens.

Less than 10 blasting events at DDMI caused PPVs greater than the DFO mandated

13 mm/s at sensor locations in the lake. During the period when the studied fish eggs

Page 20: University of Alberta Library Release Form

3

were in the lake only 3 events actually exceeded the guideline. The biological study noted

that no fish egg mortality could be attributed to the blasting at DDMI.

As a result of these findings, and in keeping with objectives 4 and 5 (listed below), a

lab study was developed where eggs were exposed to PPVs much greater than any

guidelines to try and determine a PPV that may actually increase mortality of fish eggs.

Of interest also is that guidelines for blasting all look at the maximum particle

velocity while ignoring many other components of a blast wave. Frequency, acceleration

and duration of an event are also important factors that may determine mortality of

incubating fish eggs.

This work will focus on the geophysical aspect of the project. Results from the field

project are presented as well as the results from an in lab blast simulation experiment.

Results from the biological component will be presented only briefly and not discussed.

For more information on the biological component the reader is referred to work by

Faulkner et al. (2006) and Faulkner (2006).

This report is divided into two main parts; the field component and the lab component

of the studies.

1.1) Objectives:

The objectives of the overall project based on the initial proposal submitted to DDMI,

and consistent with DFO authorisation, are as follows:

1) Determine the spatial extent of the "blast zone" beyond the dike of the A154 pit and

determine the attenuation of peak particle velocity and over-pressures away from the

centreline of the dike; as per DFO requests this will require monitoring of as many

blasts as possible both inside and outside of the predicted blast zone. Further,

Page 21: University of Alberta Library Release Form

4

measurements will assess differences between particle motion velocities and pressures

during the year with and without lake ice.

2) Monitor areas outside of the dike, regularly during the spawning season and less

intensely during routine inspections during other times, for the presence of dead or

moribund fish. Dead fish will be examined to assess the degree of trauma to internal

organs;

3) Determine if mortality of lake trout eggs is increased within the blast zone, relative to

a reference site and if it is, determine to what degree hatching success is reduced;

4) If egg mortality is increased, determine the relationship of mortality to peak particle

velocity and/or over-pressure, allowing delineation or calculation of a "realized" blast

zone;

5) Assess if enhanced egg mortality and reduced hatching is due to limited exposure

during early development or the result of extensive exposure throughout the incubation

period; and

6) Assess if geophysical and biological effects of blasting attenuate over time as the pit

deepens.

Objective 1 was covered by the ‘in lake monitoring programme’ where the PPVs of

most blasts that take place at DDMI were measured. Objectives 2 through 5 are directly

related to the biological component of the project and will not be discussed intensively

here. Objective 6 is the last phase of the studies and it is related to the deepening of the

pit.

Page 22: University of Alberta Library Release Form

5

Chapter 2) Background

Blast monitoring is a standard practice anywhere that blasting is taking place.

Protection of both cultural objects and the environment are of concern to blasters.

Monitoring of blasts allows blasters to keep the level of the blast and its associated blast

wave within safe levels such that damage to the surrounds is kept to nominal levels. This

chapter will review some of the background associated with the basic principles of

blasting, blast monitoring, and the physics behind blast (seismic) wave propagation and

attenuation.

A recent study of regulations on blast-induced vibrations (Schneider, 2001) found a

large discrepancy between regions on the maximum allowed levels of shaking,

commonly known as peak particle velocity. Where regulations exist, ground vibration is

limited to a peak particle velocity (PPV) anywhere between 5 mm/s and 50 mm/s. The

only Canada wide regulation limits vibration in fish habitat to 13 mm/s. To ensure

compliance, blast monitoring is a common practice. The energy can also enter a water

column and the resulting pressure wave, commonly referred to as overpressure, can be

harmful to aquatic animals such as fish.

Studying the blasts and measuring the resulting Peak Particle Velocities (PPVs) at

given distances for given charge weights per delay allows for a statistical analysis from

which an empirical relation is produced for future prediction of PPVs. Guidelines which

are based on measurement of Peak Particle Velocity on land and Over Pressure in the

water column are in place for protection of everything from buildings to aquatic life.

With some prior site knowledge, a simplified “back of the envelope” calculation can be

made to help blasters stay well within the localised guidelines and quickly determine the

Page 23: University of Alberta Library Release Form

6

maximum safe blast weights as will be discussed in detail in later chapters. Knowing the

densities of materials and their acoustic properties, we can model the PPVs and

overpressures in order to determine their potential effect, as is discussed in detail in later

chapters.

When energy is released in sudden, catastrophic events such as earthquakes or due

to large detonations of explosives, seismic energy propagates outwards from the source

equally in all directions and the distance a detectable seismic wave travels depends on the

nature of the medium that carries it. A variety of factors, such as seismic absorption and

scattering due to variations in density and elasticity, influence a seismic wave’s

propagation and hence its intensity at a given distance from the activating source. For

example, in an ideal, non-attenuating substrate, the intensity decays in inverse proportion

to the square of the distance from the source. At geologic boundaries, further energy is

lost from the forward propagating wave as only some energy passes through the

boundary while the rest reflects back. The reflection is dependent on the density contrast

between the two media and the angle of incidence. Here an overview of the progressive

reduction in the amplitude of a seismic signal as it travels farther from the point of origin

is provided in order to allow for the understanding and interpretation of the observations.

2.1) Basic Physical Principles of Seismic Waves

Seismic waves are transient disturbances propagating on the surface and in the

subsurface and created by any impact or failure on the surface or within the earth. The

term, seismic wave, however, is too generic and here, the types of waves, and their

behaviour, are described below.

Page 24: University of Alberta Library Release Form

7

A blast wave is composed of both body and surface waves. The propagation, or wave,

velocity is the speed at which the wave travels. As the wave-front propagates its intensity

declines due to seismic attenuation. The observed amplitudes, which are of concern here

in blast monitoring, will depend on factors such as the density and elasticity of the

material, the distance from the source to the receiver, scattering from heterogeneities, and

attenuative properties. Each of these aspects are briefly reviewed.

2.1.1) Basic Wave Propagation

When seismic waves propagate, they carry energy through the ground or along the

surface. In the context of blast monitoring, it is crucial to understand what this means. A

general ‘elastic’ wave, which includes sound and seismic waves, is essentially a transient

disruption of the material that travels at a speed c. That is, as the wave passes a given

point in the earth, the ‘particle’ at that point undergoes motion; and from basic physics

we know that we can look at this in terms of the particle’s displacement, its speed, or its

acceleration. If the particle is in a material continuum, it will also be subject to pressure

or stress. Usually these motions are considered to be ‘harmonic’ meaning that the

particle motions are described by simple sinusoidal function with a given frequency. In

one dimension ‘x’, we can define the particle displacement ux, the particle velocity ux’

and the particle acceleration ux’’ of wave propagating along the x axis with x-directed

displacement through this harmonic function and its derivatives with respect to time.

)cos( kxtAux −= ω (2-1)

)sin(.

kxtAux −−=′ ωω (2-2)

and

Page 25: University of Alberta Library Release Form

8

)cos(2..

kxtAu x −−=′′ ωω (2-3)

where ω = 2πf is the angular frequency, A is the displacement peak amplitude, k = ω/c is

the wave number, t is time and x is the location. It needs to be noted that the particle

velocity is the velocity at which the particles move when the seismic wave passes and not

the velocity of the wave front propagation. As will be necessary in later discussions in

this chapter, it should also be kept in mind that there will actually be three components of

particle motion with ( ) ( ) ( ) ( )ktxujtxuitxutxu zyx

)r)r)rrr ,,,, ++= where [ ]kji)))

,, is the set of

directional unit vectors parallel to the reference co-ordinate system [x,y,z] axes with

position in 3D space given by the vector kzjyixx)))r

++= . Similarly, the amplitude

needs to generally be described by a constant vector kAjAiAA zyx

)))r++= although much

of the discussion below will reduce this to a scalar for the sake of simplicity. The

formulae describing the particle motion in the y and z directions are similar to those for

Eqns. 2.1 – 2.3 immediately above.

Consequently, a wave can be characterised by its particle displacement, its particle

velocity, or its particle acceleration. A given receiver will measure only one of these.

For example, highly sensitive seismometers that sense the arrival of low frequency

seismic waves produced by distant earthquakes register submicron particle

displacements. Geophones are typically sensitive to frequencies used in seismic

exploration, and also in many blast monitoring applications over the band from about 10

Hz to 300 Hz and at their simplest rely on the motion of a permanent magnet through a

wire coil. The current produced in such a coil consequently depends on the speed at

which the magnet moves through the coil, according to Faraday’s Law; and the output of

the geophone relates directly to the particle velocity. Accelerometers are used in

Page 26: University of Alberta Library Release Form

9

situations where strong ground motions are expected and are used to assist in

determination of earthquake hazards; the values from these strong motion instruments are

often given as a ratio with respect to the earth’s gravitational acceleration.

As will be seen later, the instruments deployed here in blast monitoring typically

involve geophones primarily due to their simplicity and robustness. The jargon used in

the engineering blast monitoring community is that the particle velocity, which is a

function of time, is represented as ‘PV’. Further, most regulatory guidelines are given in

terms of ‘peak particle velocity’ (PPV) with units of in/sec in the United States and mm/s

in the rest of the world. Schneider (2001) provides a relatively recent review of existing

blasting regulations in different jurisdictions worldwide.

A hydrophone is the second type of transducer that is used in this thesis to record

the propagation of wave energy. Hydrophones are sensitive to the pressure of a passing

sound wave in water and as noted in the first chapter there are also guidelines on

allowable ‘overpressures’.

2.1.2) Elastic Moduli

However, in order to link these pressures to stresses in a solid material, it is

necessary to first briefly define the elastic moduli, which for purposes of this study will

be confined to:

1. The bulk modulus K = -∂P/(∂V/V) with units of N/m2 (Pascals) which is a

measure of the change in pressure (or more precisely the normal stress)

induced in a material due to a change in volume of the material. All solids

and fluids have a bulk modulus and can support a pressure.

Page 27: University of Alberta Library Release Form

10

2. The shear modulus µ = τ/γ in units also of Pascals with the shear stress τ

and shear strain γ. The shear stress can be envisaged as the ratio of the

stress pulling horizontally within a surface to the area of the surface. The

shear strain may, for small deformations, be defined as the resulting

decrease in the angle between two fibres that prior to deformation were at

right angles to one another. The shear modulus essentially describes the

change in the shape of an object due to deformation. Only solids may have

a shear modulus; fluids cannot support a shear stress. Indeed, this lack of

ability to support shear stress is one of the defining characteristics of fluids.

The Poisson’s Ratio ν = −εz/εx which is the ratio between the z-directed strain

εz to the x-directed strain εx induced by applying a uniaxial z-directed stress to a

bar of the material. Various physical restraints require -1 ≤ ν ≤ 0.5 however,

most earth materials typically1 will fall within 0.1 ≤ ν ≤ 0.4.

The elasticity of an isotropic material (i.e. one which has the same physical

properties regardless of the direction from which it is observed) is completely

described by only two elastic constants. K and µ are not the only ones that are

defined; two others are the Lamé parameters λ and µ, introduced mostly due to the

mathematical simplifications they allow. The second µ is the same shear modulus.

The first λ = K – 2µ/3 does not have a simple physical definition but in the absence

of a shear modulus µ for a fluid collapses to K. With the use of the elastic

constitutive Hooke’s law may be written succinctly as:

1 A negative Poisson’s ratio means that the width of an object will shrink when it is compressed along its length, such materials do exist but are usually only artificially produced for specific purposes.

Page 28: University of Alberta Library Release Form

11

ijijij µελθδσ 2+= ( 2-4)

Where θ = (ε11 + ε22 + ε33) is the volumetric strain, δij is the Kronecker’s delta (δij =

1 when i = j and δij = 0 when i ≠ j), and εij is the strain defined by:

∂+

∂∂

=i

j

j

iij x

uxu

21ε (2-5)

As noted, a wave is a propagating disturbance that results in both changes in volume,

shape, and stress. For a perfectly elastic medium (i.e. one in which there is no intrinsic

loss of energy due to wave absorption) one may derive a vector wave equation by

considering: 1) Hooke’s Law constitutive relationships between stresses, the elastic

moduli, and strains, 2) Newton’s second law F = ma that relates the acceleration of a

small element in the material due to unbalanced forces on either side of it as the wave

passes. Again, the details of the derivation of this equation can be considerable and are

provided in many texts (e.g. Stein and Wysession, 2003) but the equation takes the form:

( ) ( )( ) ( ) ( )2

22 ,,,

ttxutxutxu

∂∂

=∇+•∇∇+rr

rrrrrrrρµµλ (2-6)

where overhead arrows signify a vector operation of quantity, ∇r

, •∇r

, and 2∇r

are the

vector gradient, divergence, and Laplacian operators, respectively, and ρ is the mass

density. The solutions to the wave equation take a harmonic form:

( )xktieAtxurrrrr −= ω),( (2-7)

Page 29: University of Alberta Library Release Form

12

where kr

is the vector wavenumber. While this all looks very imposing, the final results

upon examination of the solutions is relatively simple for an isotropic elastic medium. In

general, two ‘body’ waves will exist: a compressional or P-wave2 with wave velocity:

ρ

µλρ

µ 234

+=

+=

KVP (2-8)

And with particle displacements parallel to the direction that the wave is propagating, and

a slower shear or S-wave3 propagating at

ρµ

=SV (2-9)

One final comment here is that VP > VS always and the theoretical limits for an elastic

medium are enforced by;

3. Poisson’s ratio is also tied to the VP/VS ratio via:

( )( ) 1/

2/21

2

2

−−

=SP

SP

VVVVν (2-10)

A part of the range of which is plotted in Figure 2-1 for illustration.

2 The P-wave is also known in other disciplines as the longitudinal wave to signify the relation between the particle motion and wave propagation direction. Compressional refers to one aspect of the material behaviour as the wave passes the volume of the material dilates and contracts. The origin of ‘P’ wave is from seismology and likely refers to the fact that this wave has the higher velocity and as such arrives first and is therefore ‘primary’. 3 This wave is also known in other disciplines as a transverse wave. The designation of ‘S’ is also not known but likely arises from its ‘secondary’ arrival or from its shear nature.

Page 30: University of Alberta Library Release Form

13

Figure 2-1) Poisson ratio plotted against the Vp/Vs ratio.

Body waves propagate through a three dimensional medium; in the current context they

will appear as the P-wave, the S-wave, and the air-blast wave. However, these are not the

only seismic waves and, indeed, are often not the ones of most concern in surface

vibration monitoring. The strongest ground motions, in earthquake and exploration

seismology and in blast vibration monitoring are generally Rayleigh4 surface waves.

Rayleigh waves dominate at large distances for blasting or construction vibrations.

(Dowding, 1985). These seismic waves exist whenever there is a ‘free’ surface, such as at

4 These waves were first described by William Strutt, third Baron Rayleigh (1842-1919) who wrote prolifically on many topics and particularly on the propagation of sound waves. He received the Nobel Prize in Physics in 1904.

Page 31: University of Alberta Library Release Form

14

the contact between the solid earth and the atmosphere, the influence of the latter can

usually be ignored. The speed of propagation of these waves VR also depends on the

elastic physical properties of the half-space upon which they travel, and they must be

found from the zero that satisfies the requirement 0 < VR < VS:

011422/1

2

22/1

2

22

2

2

=

−+

P

R

S

R

P

R

VV

VV

VV

(2-11)

The relationship between VR and the VP/VS ratio, also given as a ratio of VR/VS, is plotted

in Figure 2-2, which shows that VR is about 90% of VS.

Figure 2-2) The Vr/Vs ratio plotted against the Vp/Vs ratio.

The Rayleigh wave particle motions are substantially more complex than those for the

linearly polarised P- and S-waves and they are dependent on a variety of factors of which

the depth is the most important. Again, surface waves are trapped by and decay

Page 32: University of Alberta Library Release Form

15

exponentially away from the surface. Finding these particle displacements is laborious

and here an example of the formulae for the two components of particle displacement at

the surface for a Poisson solid (i.e. one with ν = 0.25) for a Rayleigh wave propagating in

the positive x-direction will be

( )( )xktAktxu

xktAktxu

xxz

xxx

−=

−=

ω

ω

cos62.0),(

sin42.0),( (2-12)

These formulae define an elliptical ‘retrograde’ particle path which means that the shape

of the path is elliptical but that the trajectory as it rises along the path is reverse to the

wave propagation direction as indicated in Figure 2-3.

Figure 2-3) Retrograde particle motion of a Rayleigh wave. Wave propagation direction along the surface is represented by the large horizontal arrow. At a given point the particles move ‘backwards’ around the ellipse as shown.

Page 33: University of Alberta Library Release Form

16

This elliptical motion is similar to that experienced by a small boat on the water and

which causes sea sickness, although these ‘gravity’ waves on water have the opposite,

prograde motion.

The current discussion only touches some of the basics of surface wave propagation.

The real earth is much more complex than the simple half-space considered above. For

example, another surface wave called the Love5 wave is produced by the interference of

horizontally polarized shear wave reverberations within a low velocity layer over a half-

space. Love waves are by their very nature ‘dispersive’ meaning that the speed of their

propagation depends not only on both the variations in the shear wave speeds between the

layer and the underlying half-space but particularly on the frequency. This results in

spreading of the wave form pulse with distance from the source.

The behaviour of the Rayleigh wave, too, becomes more complex even with such a

trivial addition to the geological structure. There are two new aspects to this. First, the

Rayleigh wave also becomes dispersive with its velocity now dependent on the frequency

and, second, there are additional interfering Rayeigh wave modes produced by the

interference of reverberating P- and S-waves. While such complexities complicate

analysis of Rayleigh waves, they do add considerable information that can be profitably

used in delineating the subsurface structure (e.g. Beaty and Schmitt, 2003). While these

are useful points to consider, however, they are beyond the current scope of this

contribution which focuses mostly on the monitoring aspects of blast vibrations from a

rather basic level. That does not mean that such complications should not be included to

further refine blast monitoring technology in the future.

5 The Love wave was first predicted theoretically by and is named after A.E.H. Love (1863-1940) who worked extensively in the theory of elasticity and wave propagation.

Page 34: University of Alberta Library Release Form

17

2.1.3) Wave Energy

In blast vibration monitoring it is important to understand what influences the

observed amplitudes and this first requires some idea of the energy carried by a

harmonic6 wave which is, without derivation (see Stein and Wysession (2003), pg 35 &

pg 61), given by

2

22 ρωoAE = (2-13)

where E is the average over one wavelength of the total energy conveyed per unit

wavefront (in units that reduce to J/m3 and as such is analogous to an energy density), ρ

is the density of the material, ω is the angular frequency and Ao is the peak particle

displacement amplitude. Simple examination of the last equation shows that the energy

is proportional to the squares of both the peak amplitude and the frequency. The

magnitude of the vector energy flux F is the rate at which the wave transfers the energy

past a given point and is found by multiplying by the wave velocity, hence for the P-wave

the flux F is

22

2222 ZAVAF oPo ωρω== (2-14)

Where Z = ρVp is the acoustic impedance of the material. Z is a measure of the resistance

to motion of the mateiral by the passing wave. Note that dimensionally F has units of

kg/s3 which is dimensionally the same as the intensity: the power per unit area in

Watts/m2; this is the term that is more normally used in introductory physics texts when

sound wave energy is discussed.

6 Harmonic means monofrequency, i.e. a cosine or sine wave.

Page 35: University of Alberta Library Release Form

18

The last two equations really only consider propagation of the wave within a single

medium with uniform properties. They are more instructive when the case of constant

energy flux is compared between two media with impedances Z1 > Z2 which upon

examination of the last equation shows that the peak amplitude in the first medium A1

must be less than that in the second medium A2. That is, a dense and high wave speed

material is displaced less than a lighter and low wave speed material by a passing wave of

the same intensity.

The last two equations allow relation of the wave energy to the particle displacement, and

hence to particle velocities. One other aspect of the current study, however, is the

measurement of pressure variations within the water column due to blasting. The

pressure P, or more precisely the variation of pressure from the equilibrium atmospheric

in air or hydrostatic in water, of the wave is coupled to the particle motions. Later with

regards to blast monitoring this will be called the ‘over-pressure’. As mentioned above,

in fluids only a P-wave is transmitted and as is well known such ‘sound’ or ‘acoustic’

waves have a particle displacement in the direction of the wave propagation and are

related to other characteristics of the wave propagation within:

FZAFZA

AZAZAVPo

ooooP =

′=

′′=′===

ωωωρ (2-15)

where oA , oA′ , and oA ′′ are the peak particle displacements, particle velocities (‘PV’

above), and particle accelerations due to the passing wave, respectively.

Page 36: University of Alberta Library Release Form

19

2.1.4) Decay of Amplitude

Understanding the factors that influence the observed amplitudes is critical in the

interpretation of blast vibration records. Energy loss, or decrease in the amplitude, occurs

as the waves pass through media. Seismic waves lose energy through absorption,

reflection and refraction at interfaces, through mode conversion and through spherical

divergence, or spreading of the wave. The primary factors crucial to decay of the seismic

signal are geometrical wavefront spreading, intrinsic attenuation, and scattering

attenuation. More details on this may be found in Grech (M.Sc. thesis, 1998).

Perplexingly, the blast vibration literature has focussed primarily on highly empirical and

statistical descriptions without much in the way of a critical examination of the

underlying physics of the problem, a deficiency recently discussed by Blair (2004). This

is likely because, as will be hinted at below, the decay of surface waves is relatively

complicated. Below, I re-examine briefly some of the important aspects that will

influence the decay of amplitudes from the seismic source7.

i) Geometrical Spreading

The above discussion has focussed on the propagation of monofrequency ‘plane

waves’: an idealized theoretical construct that simplifies mathematical descriptions of

wave propagation. As the name suggests, a plane wave front is flat and two-

dimensional; this can be useful if one is in the ‘far-field’ of a seismic source where the

wavefront is for practical purposes flat as, for example, studies of ‘teleseisms’ in global

seismology.

7 A seismic source is considered to be a controlled source used in exploration. The source may be an explosive, a hammer or a seismic vibrator.

Page 37: University of Alberta Library Release Form

20

This assumption is not adequate in the current context of blast vibration and over-

pressure monitoring where the source can be within tens of metres of the recording

sensor. The simplest approximation is for body waves propagating from a hypothetical

‘point’ source in a three-dimensional continuum. In such a case the wave energy moves

out in all directions radially from the source with the energy of the wave uniformly

distributed over the surface of a sphere. Of course, the radius R of the sphere increases

linearly with time at a rate proportional to the wave speed. Correspondingly, the surface

area of this expanding sphere increases in proportion to R2, and since energy is conserved

on the wavefront the intensity must decrease in proportion to 1/R2. Consequently, the

observed amplitudes, regardless of whether they are for particle displacements, velocities,

or accelerations, will decay proportion to 1/R.

The situation differs somewhat for surface waves that propagate over a two-

dimensional surface. A circular surface wave front also expands out along the surface

proportionally to the wave speed but the energy is now distributed along the circular

locus and hence the local intensity at a point on the wavefront decays less rapdily at 1/R

dependence. Similarly, the amplitudes have a 1/√R dependence. This slow decay of

amplitude is the major reason why surface waves are so strong relative to either P or S

waves. In summary, the amplitude decay is proportional to r/1 for a body wave and

r/1 for a surface wave; and in practice Rayleigh waves will dominate typically within

only a distance of 2 wavelengths from the source (Lai et al., 2002).

In general, the rate of decay of the particle velocity due to the divergence of the wave

front can be described by

Page 38: University of Alberta Library Release Form

21

n

RRyy

=

2

1.

2

.

1 (2-16)

where .

1y and .

2y are peak particle velocities at radii R1 and R2, n is 1 for a body wave

except at surface where n is 2 and, for a Rayleigh wave, n is 1/2.

Dispersion, which is particularly important in the propagation of surface waves, has

another effect that is related to the geometry of the situation. As noted above, the

duration of a dispersive seismic pulse lengthens with propagation distance. This means

that the energy of the pulse is spread out over a longer time; the longer the pulse travels

the smaller the instantaneous intensity and hence the instantaneous amplitudes will be.

While this may be obvious to a seismologist, this factor does not appear to be considered

in the design of the somewhat arbitrary predictive empirical formulae for blast vibrations

as will be discussed later, although engineers should not be faulted as this remains an area

of active scientific study. This is particularly true in the near surface geophysical

community where there have been more recent attempts to examine the truer geometrical

decay of Rayleigh waves using cylindrical waves described with Bessel functions instead

of by plane waves (Zywicki and Rix, 2005). This recent work shows that the square root

decay provides an estimate of the minimum decay experienced by the Rayleigh wave.

ii) Intrinsic attenuation

The energy contained by the wave and its amplitude attenuate as they pass through

media. The physical properties of the media dictate the rate of attenuation. All materials

will absorb seismic energy to some degree, and because this absorption is irreversible it is

often also referred to as internal friction. There are a variety of mechanisms that have

been proposed for this absorption that in include thermally dependent motion of defects

Page 39: University of Alberta Library Release Form

22

in crystals and grain boundary sliding (applicable in the deeper crust and mantle and in

metals), frictional sliding on mineral grains (in weakly consolidated rocks), and

differences between the motion of the saturating fluids and the solid mineral frame due to

the wave passing.

The attenuation results in the progressive dissipation of energy as the wave travels

also contributing to the loss of amplitude. That is, one form of energy is converted into

another as the wave travels. In this case, example, seismic waves are partially converted

to heat as they pass through rock.

iii) Attenuation at geologic boundaries

At physical boundaries, or geological interfaces energy is lost as only some of the wave is

passed into the next medium while the rest is reflected. The percentage of reflection is

dependant on the density contrast between the media. Absorption, or material damping is

the energy lost due to friction. With each cycle (wavelength) the blast wave loses a small

amount of energy to overcome friction. It is defined by

)(.

1

.

212 RReyy −−= α (2-17)

where α is the attenuation coefficient which is given in units of m-1.

The attenuation co-efficient is frequency dependent for a given material and one way to

overcome this limitation is with the quality factor Q is the ratio between the total energy

stored by the wave to the energy lost per radian (or equivalently 2π times the total energy

stored by the wave to the energy lost per cycle). So, the less energy lost per cycle the

greater the value of Q. Q is related to α via:

oo

o

o

o

o

oo V

fV

VV

Qαπ

αω

ωα

αω

ω =≅

−=

21

2)( 2

22

(2-18)

Page 40: University of Alberta Library Release Form

23

with the right-most approximation essentially active under the assumption that Q remains

constant. Hence, because Q is usually assumed to remain constant or nearly so it more

conveniently describes the attenuation than a. For the sake of comparison, typical values

of Q are >106 for fused quartz glass, ~105 for aluminium metal, ~ 50 for P-waves and

~30 for S-waves in acrylic glass (e.g. Plexiglas™), ~ 200 or more for typical

metamorphic rocks, ~ 40 or more for sedimentary rocks, and as low as ~15 for bitumen

saturated sands.

iv) Scattering Attenuation

A third type of amplitude decay is caused by the ‘scattering’ of the seismic wave by

elastic heterogeneities such as pores, fractures, and any other features of the geology that

result and spatial variations of the wave speeds and densities. In scattering attenuation

no energy is dissipatively lost by conversion to heat as for intrinsic attenuation above, but

instead portions of the energy are partitioned between the forward travelling ‘ballistic’

wave pulse that propagates from the source to the receiver and the ‘scattered’ energy that

moves off in different directions. The simplest form of this scattering is the division of

the energy of a pulse by simple reflection and transmission at a geological boundary.

The effect of this is that the energy remaining in the forward moving ballistic pulse, and

hence its amplitude relative to the total energy in the system, is diminished every time

such a boundary is encountered.

Scattering attenuation is a large, complex, and still intensively studied topic.

Scattering of light is experienced everyday, however, and this provides some frame of

reference for discussion of the topic. For example, the sky is blue because of the

Page 41: University of Alberta Library Release Form

24

Rayleigh8 scattering of the sun’s light by the small molecules in the air. The degree of

scattering is proportional to 1/Λ4 where Λ is the wavelength; therefore, the bluer and

smaller wavelength fractions of the white sunlight are preferentially scattered back to us

to make the sky blue. When one looks more directly at the Sun, however, the light

surrounding it does not appear to have any colour, this is because of Mie9 scattering from

larger particulates, which is not so frequency dependent. Seismic waves are similarly

subject to these kinds of scattering.

It is difficult to separate the effects of Rayleigh and Mie scattering from the intrinsic

scattering discussed earlier. A practical outcome of this is that these effects may be

incorporated into an experimental determination of Q.

v) Combination of effects

Combining equations 2-4 & 2-5 we can define the attenuation of the amplitude from y1 to

.

2y as the wave produced by a point source propagates out from radial distance R1 to R2

can be generally be described by:

)(

1

2.

1

.

212 RR

n

eRRyy −−

= α (2-19)

where again n = 1 for body waves and n = ½ for surface waves. Again, however, it must

be noted that this equation is based on rather simple assumptions and workers are still

actively working on more physically meaningful solutions to this complex problem of

8 Again, first considered by the same Lord Rayleigh discussed above. In Rayleigh scattering the wavelengths are much larger than the scattering particles which in the case of the blue sky are mostly N2 molecules. 9 Mie theory is named after G. Mie (1868-1957) who first developed the theory of the scattering of an electromagnetic wave (e.g. light) by a dielectric (insulating) sphere in 1908. In Mie scattering the wavelengths are roughly comparable to the particulate size.

Page 42: University of Alberta Library Release Form

25

Rayleigh waves (e.g. Lai and Rix, 2002; Zywicki and Rix, 2005). In order to gain a

better understanding of this equation, the decay of amplitude with distance from the

source to the receiver is plotted in Fig 2-4 below. This case uses some parameters that

will be realistic for the blast monitoring case studied with VP = 5000 m/s and f = 100 Hz.

Assuming a uniform reference intensity with value of unity at a reference distance of 10

cm from the center of the radiating sphere (for body waves) or cylinder (for surface

waves) the loss of amplitude is calculated for three cases of Q = ∞, 100, and 10 for the

surface wave for comparision10. This shows the substantially larger difference in the

geomtric decay between surface and body waves; for the same initial amplitude at the

reference distance here of 10 cm, there is a 40 dB (100 times) difference in their

amplitudes at 1 km offset. This further highlights the relative importance of surface

waves in controlling the maximum amplitudes.

10 Fig.2-4 is plotted in terms of units of dB (decibels) versus distance which is an accepted standard measure of relative amplitude, this is named in honor of the inventor of the telehone, Alexander G. Bell. Because of the square relationship between amplitude and intensity, the relative value plotted is 20log10(y/yref) . Consequently, a signal that has decreased by 10 dB is 100.5 = 3.16 times smaller amplitude, 20 dB is 10 times lower, 30 dB is 31.6 times lower, etc. It must be remebered that the usage here is in terms of particle amplitude, whether it be displacement, velocity, or acceleration. This is similar to, but differs somewhat, from the dB scale most readers will be familiar with from knowledge of sound intensity with a well defined reference intensity.

Page 43: University of Alberta Library Release Form

26

Figure 2-4) Illustration of the relative decay of a cylindrically spreading surface wave and a spherically spreading body wave with distance relative to the (arbitrary) reference distance of 0.1 m for a wave speed of 5000 m/s and a frequency of 100 Hz. Relative amplitudes report in dB10. Numerical values represent quality factors included in the calculations. Q = ∞ means no attenuation.

An important question to ask is what is the actual displacement that might be

expected from an explosion? The total amount of energy released by the explosive

material itself can be taken to depend linearly on the amount of explosive used and is

usually calculated relatively easily from knowledge of the chemistry. For example,

Trinitrotoluene TNT (C6H2(NO2)3CH3) releases approximately 616 kJ/mol or 2175 J/g.

However, what is just as or more important than the amount of energy released are other

factors such as the ‘detonation velocity’ which for TNT is nearly 7000 m/s11. Even this is

a gross simplification, as the detonation velocity will depend on other factors such as the

11 Surprisingly, gasoline releases much more energy but is not as explosive (but still dangerous) and only has a flame advance speed of 0.34 m/s in air.

Page 44: University of Alberta Library Release Form

27

size of the hole, the type of explosive, the peak pressure produced, the shape of the

charge, and the strength of the rock surrounding the explosive.

The ultimate energy of the explosive is converted directly to heat, to fracturing,

crushing, and moving the rock, and to seismic energy. The fraction of this available

explosive energy that actually goes into producing seismic waves (vibrations) is very

difficult to determine. Hinzen (1998) carried out a number of simple tests in a hardrock

(gneiss) mine that showed the seismic efficiency (i.e. the percentage of energy radiated as

seismic waves to the chemical energy of the explosive) to range from 2% to 25%

depending on the geometry of the source. However, this applies more to near field body

waves than to the more typical surface situation where the Rayleigh waves will be

important.

In any case, despite all these complications, the amount of energy in seismic waves

depends on the total amount of explosive detonated. This is usually given as the ‘blast

weight per delay’ which is the total amount of explosive detonated at ‘the same time’

with the same time usually defined within an 8 ms interval.

2.2) Introduction to Blast Monitoring

In field practice, effects on the environment are measured as the peak particle velocity

(PPV) for ground shaking and over-pressure for energy transferred into the water

column. As noted in the previous section, PPV will be measured by the maximum value

of the magnitude of the vector particle velocity as will be described below; for illustration

however the PPV value is shown in Fig 2-4.

Page 45: University of Alberta Library Release Form

28

Figure 2-5) Illustration of the definition of the PPV on a particle velocity versus time plot where the peak particle velocity is the maximum excursion from zero amplitude.

Different levels of each are deemed acceptable for different industries and different

parts of the world (e.g. Schneider, 2001) but there appears to be little in the way of

consistency in the way many of these regulations have been developed, measured, or

applied. Grogan and McAnuff (2005) recently provided a short review of the blasting and

vibration regulatory environment in Canada from the perspective of mining and

construction; they found that that most provinces have no set regulations. Ontario and

Nova Scotia require routine monitoring if the PPV levels exceed 12.5 mm/s while

Manitoba sets this level at 12 mm/s. In Alberta and British Columbia, mine inspectors

evaluate each case on an individual basis and the limits become part of the permit. In

contrast, Alberta, British Columbia, and Saskatchewan all have restrictions on seismic

exploration. These are not described by PPV but are given in terms of the distance that

the seismic source, which can also be a seismic vibrator instead of explosives, must be

from a particular feature such as a pipeline. For the sake of comparison, the current

Alberta government (2005) regulations are reproduced in Table 1. The basis of these

Page 46: University of Alberta Library Release Form

29

regulations is not known but it is likely that they originated from a series of studies in the

late 1970’s on the influence of seismic blasting on water wells (Goble, 1980)

Table 1) Current Alberta seismic exploration regulation offsets, Schedule 2 of Alberta Government (2005).

Explosive

Structure Charge Weight (kg) Dist. (m)

Non-Explosive Dist. (m)

Test Hole Dist. (m)

2 or less 64

Greater than 2, not greater than 4 90

A. Residence, barn, cemetery, a building or

Greater than 4, not greater than 6 110

structure with a concrete base, irrigation

Greater than 6, not greater than 8 128 50 64

headworks, dam, water well Greater than 8, not greater than 10 142

Greater than 10, not greater than 12 156

Greater than 12, not greater than 20 200

Greater than 20, not greater than 40, 284

Greater than 40, not greater than 100 450

2 or less 32

Greater than 2, not greater than 4 45

Greater than 4, not greater than 6 55

B. High pressure pipeline (measured)

Greater than 6, not greater than 8 64

From the centre line of the pipeline, oil or gas well

Greater than 8, not greater than 10 70 15 32

Greater than 10, not greater than 12 78

Greater than 12, not greater than 20 100

Greater than 20, not greater than 40 142

Greater than 40, not greater than 100 225

C. Driveway, gateway, survey monument, buried telephone or telecommunications line

All 2 2 2

D. Low-pressure distribution line (measured from the centre line of the pipeline)

All 3 3 3

Page 47: University of Alberta Library Release Form

30

E. Irrigation canal more than 4 metres wide

All 10 10 10

F. Buried water pipeline All 10 10 10

The only Canadian wide regulations are enforced by the Department of Fisheries and

Oceans (DFO). Wright & Hopky (1998) state that maximum allowable PPV in fish

habitat is 13 mm/s and maximum allowable pressure in the water column is 100 kPa

(approximately 1 atmosphere of overpressure). DFO provides sample calculations for

calculations of the PPV and water pressure based on the charge weight and the distance

between the source and the receiver (Eqns. 2-22 to 2-25).

Much literature is devoted to blast monitoring for prediction of damage to cultural

items. Effects on aquatic life from on land blasting is only discussed by regulating bodies

such as the DFO. Monitoring of blasts at construction, mining or any other sites where

activities induce ground vibration is common practice.

Blasts are monitored to measure and predict the effects on cultural and natural

objects, for example; fish habitat or man-made structures such as dikes. Ideally, the

monitoring and prediction will keep effects of the blasting below damaging levels.

2.2.1) Peak Particle Velocity

The Peak Particle Velocity is the maximum particle velocity attained when the

seismic wave passes a measurement point on the earth (Figure 2-5). In most of the world

its values are generally noted in metric units of mm/s or cm/s. As noted above, it should

be remembered that this is the velocity at which the particles are moving and not the

speed at which the wave front is moving. The actual wave front velocity is of no

importance to this project as we are only interested in the energy of the wave front and its

Page 48: University of Alberta Library Release Form

31

attenuation. In Canada, as per DFO regulations, the maximum allowable shaking in the

substrate is 13 mm/s. Shaking beyond this level is considered to be potentially harmful to

incubating fish eggs.

Particle velocity is measured using three orthogonally placed geophones usually

oriented with respect to the vertical and horizontal. The observed PPV is calculated from

the responses of the three geophones SH1(t), SH2(t), and SV(t) that respectively represent

the time dependent particle velocities for the two horizontal and one vertical components.

Referring back to Eq. 2-7, these explicitly would look like:

( ) ( ) ( ) ( ) ( ) ( )( )ktSVjtSHitSHkAjAiAeeAtxu zyxxktixkti ˆˆˆˆˆˆ),( 21 ++=++==′ −−rrrrrrr ωω ωω (2-20)

These may be taken as components of the particle velocity vector with magnitude

with time PV(t) given by:

222 )()(2)(1)( tSVtSHtSHtPV ++= (2-21)

with PPV equal to the maximum value of PV(t) over the time window of interest. This is

a simple attribute extracted from the data and it is worth noting that while this is a useful

measure it does not incorporate any information with regards to the direction of

maximum shaking, its duration, or its acceleration. It is also important to note that this

definition is often not consistent within the engineering literature or even within existing

regulations from place to place; often, only the vertical component SV is used, or in other

cases, the peak-to-peak value may be used.

To predict the peak particle velocities, DFO provides a simplified calculation where

the minimum safe distance from the blast is calculated based on the maximum charge

weight to limit the PPV to 13 mm/s or less.

09.15*WR = (2-22)

Page 49: University of Alberta Library Release Form

32

where R is the distance from the blast location in metres and W is the charge weight per

delay, in kilograms, of the blast.

The PPVs can be modelled once several blasts have been monitored based on

statistical methodology where W

R is considered to be the scaled distance.

2.2.2) Overpressure

In the blast vibration monitoring lexicon, overpressure is defined as a pressure

deficit/excess created by the seismic pressure wave in the water column and is measured

in kiloPascals (kPa)12 using what is known as a hydrophone. In Canada, as per DFO

regulations, a pressure change over 100 kPa is considered to be harmful to aquatic life.

To approximate a minimum offset for a given charge weight a series of

calculations need to be made as per DFO. Firstly, the transfer of pressure from the

substrate to the water column is calculated, the pressure in the substrate is found using a

maximum value of 100 kPa in the water column (Equations 2-23 & 2-24).

)/(1

)/(2

rw

rrww ZZ

PZZP

+= (2-23)

rr

wwrw CD

CDZZ = (2-24)

here Pw is the pressure (kPa) in the water column, Pr is the pressure in the substrate and

Zw and Zr are the acoustic impedances of water and the substrate respectively, Dw and Dr

are the densities (g/cm3), Cw and Cr are the compressional wave velocities. Then the PPV

due to pressure is calculated

12 For reference, one standard atmosphere of pressure is 101.325 kPa.

Page 50: University of Alberta Library Release Form

33

V

PPVP×

=ρ2

(2-25)

where P is the pressure in the substrate, ρ is substrate density and V is seismic wave

velocity in the substrate. Knowing the PPV at the substrate interface, the safe distances

and charge weights can be solved for as outlined in the previous section.

Of course, as with the PPV, once some data are in place a statistically based

empirical formula is used for future blast level predictions.

2.2.3) Scaled Distance

As may be gleaned from the discussions above, while at its base the physics of the

problem of wave decay seems fairly straightforward, there are many complications

related to the source blasting configuration; the seismic efficiency, the effective material

properties at the receiver site, and the dispersive nature of Rayleigh waves to name a few,

that make predicting vibration levels problematic. As a result, more empirical routes are

taken.

This problem has been addressed for many decades in the earthquake seismology

community who initially developed criteria to place some quantitative meausure on the

size of earthquakes. The most famous example of this is the logarithmic Richter

Magnitude Scale, which, despite its misappropriation in the media, applies to the specific

case of consistent measurements of amplitude using a specific Wood-Anderson

seismometer over the Caltech array in Southern California. As such this is more

properly called a Local Scale and it explicitly takes the form:

( ) ( ) 92.28log3)(log 1010 −∆+= tmmAM L (2-26)

Page 51: University of Alberta Library Release Form

34

where A(mm) is the peak particle displacement of the shear wave produced by the

earthquake and recorded on the Wood-Anderson seismometer and ∆t is the S-P arrival

time difference which is essentially a rough measure of how far the earthquake was from

the seismometer vault. ML = 0 was set for a 1 µm particle displacement for a 100 km

distant earthquake. So, while this may not be an elegant solution it does have buried in it

information with regards to the energy of the earthquake (via the amplitude) and the

distance of the earthquake (via ∆t). A similar scale exists for surface waves, which may

be the only phase observed, at teleseismic distances.

The blast vibration community has the inverse problem to earthquake seismologists in

that they want to predict the vibration levels generated in blasting and they usualy use an

empirical scaled distance approach. Normalisation of the blast weight per delay and the

distance from the blast zone to the sensor is known as Scaled Distance. Scaled distance

plotted against the PPV for each blast is used to create an empirical relationship which

can be used to predict blast effects at given distances for given charge weights. Blast

monitoring and predicting effects of blasting are of great importance to any operation

where blasting is used. An empirical relation is based on measuring blast induced

vibrations at different distances from the blasts for different charge weights. Essentially,

these formulae attempt to account for the decay of the seismic energy with distance and

the differing amounts of energy released during the blast. The former depends on factors

discussed in detail above such as geometrical spreading of the seismic waves which

results in lower intensity and other factors such as attenuation of the seismic waves

within the rock mass. The latter relates energy released to the mass of explosive

employed.

Page 52: University of Alberta Library Release Form

35

The blast weight per delay and distance from the blast to the sensor are first

normalised and then plotted against the recorded PPV. The scaled distance SD is:

rWDSD = (2-27)

Where D = distance, W = charge weight per delay and in practice the exponent is

typically assigned either r = ½ (Square root) or 1/3 (Cube root). The exponential factor

of 1/2 < r < 1/3 are empirical limits given in the literature. The derivation of these

exponents is not exactly clear in the literature. Dowding (1985) refers to earlier work

technical reports; but the actual development does not seem straightforward.

Apparently, these were derived not so much with some physical insight into the problem

but from an application of dimensional analysis. From a physical basis, the exponent r =

½ might be somewhat justified in that the weight of explosive detonated at any one time

should be directly proportional to the amount of energy released. As the equations above

show, the square of the amplitude of a wave is directly proportional to its energy density;

or alternatively the square root of energy density is proportional to the particle motion

amplitude. This plot is an exponential function where the PPV decreases as the scaled

distance increases. A ‘best fit’ straight line via linear regression is determined after taking

the logarithm of both axes. This best fit line is in the form Y = Mx + B and its slope and

intercept are used in the prediction function

)ln()ln()ln( ASDMPPV +×= (2-28)

where A = Y intercept giving a prediction formula for the decay of the wave with scaled

distance:

MSDAPPV )(= (2-29)

Page 53: University of Alberta Library Release Form

36

This best fit line to Eqn. 2-28 provides the coefficients for predicting the mean PPV

value. Statistical confidence limits give upper and lower bounds within which 95% of the

data fall. The upper bound is generally then used to predict safe distances from a blast.

Dowding (1985), again referring back to earlier technical reports, suggests that at short

scaled distances M ~ -2.8 while at greater scaled distances M ~ -1.6. Of course, one

problem with this is that it is difficult to actually predict what the displacements might be

at the beginning of a given project as these will be, to a large degree, site dependent. It is

only after some blasts have been made that one is able to begin to generate an appropriate

predictive curve.

2.3) Summary

This chapter has shown that the motion of seismic waves through any media is very

complicated. The effects of these waves at distances from the source are highly

dependant on many variables. Nonetheless, blasters employ empirically derived

techniques to predict shaking and to minimise the effects. At Diavik Diamond Mines Inc.

a blast monitoring programme is employed to study the effects of blasting on fish habitat.

Page 54: University of Alberta Library Release Form

37

Chapter 3) Diavik site overview

The Diavik Diamond Mine Inc. (DDMI) mine site is located on a 20 km2 island in

Canada’s Northwest Territories at 64º 30’ N latitude, 110 º 20’ W longitude some 300

kilometres north-east of the capital city of Yellowknife (Figure 3-1). Diavik Diamond

Mines Inc. (DDMI) is a world class diamond deposit with an estimated 29.8 million

tonnes of kimberlite ore averaging 3.9 carats per tonne (Diavik Fact Book, 2005) of high

grade diamonds. DDMI began production in January of 2003.

Diamonds are found almost exclusively within kimberlitic pipes: remnants of

volcanoes of which the magma originates deep within the crustal cratons. While many

kimberlite pipes have been discovered world wide, only a small percentage of kimberlites

are economically diamondiferous. DDMI has discovered four of these pipes to date at

Lac de Gras. Even though this is rich kimberlite ore, many tonnes of material must be

extracted to produce even a small amount of diamonds. The kimberlite and the host

granitoid rock is sufficiently strong that explosives must be used to break it up so that it

may be moved and processed. It is the blasting associated with this comminution of ore

and associated country rocks that is of concern because of the potential impact on the

environment of Lac de Gras.

The motivation for this study is, of course, to monitor the vibrations in support of a

larger project to study the blast effects on fish habitat within Lac de Gras itself. The lake

is 60 kilometres long and on average 16 kilometres wide. Average depth is 12 metres

with a maximum depth of 56 metres. The shoreline measures 740 kilometres. Water

temperature ranges from 0°C to 4°C in winter and 4°C to 18°C in summer (DDMI).

Page 55: University of Alberta Library Release Form

38

This chapter contains more specific background information on the site of the study

including a brief overview of the geology and the mine.

3.1) Geological Setting

Geologically speaking, Lac de Gras and DDMI’s mining operations are located in the

Archean1 central Slave Province within the Canadian Shield. The Slave Craton broadly

consists of a variety of metamorphosed granitoids, turbidites, and greenstone belts with

ages which have been dated by lead isotope techniques to be approximately 2.6 Ga, a

starting point for further investigations of this is Davis et al. (2003). In fact, the world’s

oldest remaining rock, the 4.03 Ga Acasta gneiss, lies only a few hundred kilometres to

the west of Lac de Gras. The lithosphere, or the more rigid plate-like uppermost part of

the earth, is believed to be thicker and colder than in younger geological provinces; this

colder regime may allow for both the survival of the diamonds as they are transported

relatively rapidly from great depths of ~ 200 km or more and bring the diamond stability

field closer to the earth’s surface. It is interesting to note that the diamonds themselves

are old, likely ~ 3.3 Ga, and hence have been ‘stored’ beneath deep cold lithosphere until

they were transported to the surface in igneous eruptions. In contrast, the kimberlite ‘ore’

material that brought the diamonds to the surface at Lac de Gras erupted recently during

the Eocene2 about 55 Ma.

According to publicly available geologic descriptions of the site provided by DDMI

(Bryan and Bonner, 2004), the country bedrock consists of typical 2.6 Ga greywacke-

1 The Archean Eon covers the earliest development of the earth’s crust and atmosphere as ‘recorded’ in the rock record. 2 The current boundaries of the Eocene are set at 54.8 Ma to 33.7 Ma.

Page 56: University of Alberta Library Release Form

39

mudstone metaturbites, tonalite-quartz diorites, and granites through which the kimberlite

deposits intruded. These Archean ‘basement’ rocks are also intruded by a series of

Proterozoic3 diabase dyke families. In short, aside from cross-cutting dykes, all of the

country rock is highly metamorphosed and are considered ‘hard-rocks’ or ‘crystalline’

rock. Such rocks will be typified by low or vanishing porosities, by P-wave velocities of

about 6 km/s, and S-wave velocities of about 3.5 km/s. These velocities should only be

taken as rough estimates as the wave speeds will depend on many factors such as density,

mineralogy, and texture. Schmitt et al. (2003) provides a recent compilation of the

velocities found in such hard rock.

Figure 3-1) Simplified geology at Diavik. The blue background is Lac de Gras, DDMI’s four kimberlie pipes (A154N, A154S, A418 and A21) are all situated within the waters of the lake very close to East Island where the operations are situated. (Modified from Stubley Geoscience, 1998)

3 The Proterozoic eon saw the rise of the first multi-cellular life forms and currently has the boundaries from 2.5 Ga to 543 Ma.

Page 57: University of Alberta Library Release Form

40

The surface materials at Diavik, however, differ substantially from the crystalline

bedrock. These materials are primarily glacial tills4 left about 8.5 ka by the retreat of the

laurentide ice sheet of the last ‘Tioga’ phase of the Wisconsin glaciation5. These

materials cover most of the area but they are not that thick. In the portion of the lake that

was dewatered for mining purposes the till ranged in thickness from 1 m to 3 m. On the

dry surface there is only a discontinuous layer of mixed boulders and clay-rich till

material about 30 cm thick. In the author’s on-site experience, however, many spots bare

metamorphic rock is exposed right to the surface.

As the mining operations are located in close proximity of a lake and its fish habitat,

extra care must be taken in order to ensure that exploitation activities have a minimum

impact on the fish habitat. As the preferred method of excavation of the diamondiferous

kimberlite pipes is explosive blasting creating a chemical reaction which produces a high

pressure and high temperature gas (Lucca, F.J. 2003). The high pressure and associated

energy not only breaks the targeted rock mass apart but also dissipates into the

surrounding rock.

In order to ensure the fish habitat is adequately protected the Department of Fisheries

and Oceans (DFO) has provided simplified formulae, as outlined in chapter 2, in order to

predict the maximum charge weights and minimum distances that would allow for

minimum impact on the fish habitat

As part of its commitment to minimise impact of exploitation activities on the fishes,

DDMI did its own calculations for safe standards. These calculations predict a larger

4 Till generally refers to any mixed rock-clay mixture that is directly of glacial origin. 5 The Winconsin glaciation is the name for the last major advance of the North American continental glaciers. This consisted of three phases of glaciation: the Tahoe (which peaked ~ 70 ka), the Tenaya (less well defined), and the Tioga (which peaked about 20 ka and retreated about 12 ka to 8.5 ka).

Page 58: University of Alberta Library Release Form

41

blast zone (blast zone is defined as the area within which it is expected to get PPVs

greater than the allowable ‘safe’ value) than that predicted by the DFO formulae. In turn

the predicted levels for over pressure in the water column are all substantially lower than

the allowed limit which can be shown using several different approaches.

3.2) Study site

To date four diamondiferous kimberlite pipes have been identified in this area. All of

these pipes are under the waters of the lake likely because the ultramafic kimberlite rock

weathers rapidly under surface conditions and was preferentially eroded out by the

glaciations relative to the surrounding metamorphic rock. To access the kimberlites,

DDMI undertook an immense, award winning, engineering project to build dikes around

the known kimberlite locations. Dewatering of the 154 South and 154 North pit area

(Figure 1-1) was completed in 2002. During dewatering, all fish were removed and

transferred into the open lake waters.

Mining is initially being conducted as an open pit operation up to a depth of about

400 metres at which point it will be changed to an underground operation. The first

monitored excavation blast occurred on August 27 of 2002.

3.3) Geophysical site characterisation

To aid in characterisation of the rock properties at DDMI two types of seismic

surveys were conducted; one passively recorded a mining blast in the pit while the second

was a standard refraction survey using a hammer as a source.

Page 59: University of Alberta Library Release Form

42

Figure 3-2) Figure showing the location of the seismic line receiver array and the source.

Both of these surveys were completed using the same geophone layout with the 20

receivers at a 2-m spacing were placed north-east of the pit (Figure 3-2).

3.3.1) Refraction seismic survey

The refraction survey was done using a hammer and plate as a source and the twenty

geophones as receivers. Five shot points were used spaced at approximately 5 m.

Maximum offset was at 0.5 m. The sampling rate was 4000 samples per second (fn =

2000Hz); a half second record was collected from 10 stacks. The frequency content is

very high (Figure 3-3) which is consistent with the tight spacing and dense rock type.

Page 60: University of Alberta Library Release Form

43

Figure 3-3) Frequency spectrum of refraction seismic record. The record contains, seismically speaking, very high frequencies.

Analysis of the data is difficult. Offset was insufficient to be able to see a refracted

wave and the high velocity combined with tight spacing made discrimination of wave

types impossible. Only an estimate of the surface wave is made to be in the range of 1000

m/s (Figure 3-4).

Figure 3-4) A shot from the refraction survey. Surface wave velocity is in the range of 1000 m/s. Data is displayed after ormsby filtering (10, 15, 600, 800) and RMS scaling.

Page 61: University of Alberta Library Release Form

44

3.3.2) Passive seismic survey

Two blasts were recorded but only one of the blasts was important enough to be

recorded in the DDMI data, as such, blast related data such as its location and weight is

only known for the February 12, 2004 blast. The seismic record was collected by

manually triggering the software to produce a 65 second trace sampled at 250 samples

per second. Still, at this sampling rate it was difficult to accurately resolve the first break.

An intermediate sized blast, 660 kg per delay centred about 1000 m from the

seismic array, was detonated and recorded by the geophone array. Since the exact timing

of the blast is not known we can only study the difference in time as the blast wave

reaches each geophone. Using the time difference between arrivals and the distance

between the geophone and the centre of the blast pattern we can estimate a rayleigh wave

velocity of around 2600 m/s (Figure 3-5). The frequency spectrum is much more in line

with standard seismic frequencies (Figure 3-6).

Figure 3-5) The first arrival from the blasting event. A best fit line defines the velocity to be around 2600 m/s.

Page 62: University of Alberta Library Release Form

45

Figure 3-6) Frequency spectrum for the passively recorded blasting event.

3.4) Predicted Blast Zone

DDMI (1998a.) outlines the potential for damage to both aquatic life and the dike

structure due to blasting associated with the mining operations. Sample calculations were

made to estimate the intensity and strengths of the blast wave at the dike wall and also in

the substrate of the lake and into the water column.

According to DFO guidelines (Wright, and Hopky, 1998), a maximum of 13 mm/s

peak particle velocity in fish habitat or 100 kPa in the water column is permitted. It is

unknown how these criteria were exactly derived and they, particularly for the case of the

particle velocity, are believed to be conservative estimates. However, it should be noted

that 100 kPa is nearly 1 atmosphere of pressure.

Assuming minimum distances from the outside of the dike and the blast weights that

would be used some simple predictions can be made for peak particle velocity in the

Page 63: University of Alberta Library Release Form

46

substrate and also the transfer of energy into the water column, measured as overpressure.

The Department of Fisheries and Oceans also provides simple calculations for prediction

of the PPV and over pressure. Furthermore, other calculations can be made to predict the

transfer of energy into the water column.

To calculate estimates of the effects of overpressure and PPV outside of the dike

walls DDMI made several assumptions (DDMI, 1998a) that included a typical explosive

weight per delay of 1140 Kg, a closest distance between a blast and the outer edge of the

dike wall of 310m, and the density of the till and dike material are 1.92 g/cm3 and 2.04

g/m3, respectively.

3.4.1) Peak Particle Velocity

DDMI calculated that PPVs could exceed 13 mm/s up to 400 m beyond the dike wall

and into the lake. The DFO formula gives a less conservative estimate of only 200 metres

beyond the dike. Even within the DFO blast zone, however, numerous fish spawning

habitats are found.

Using a generalised empirical formula provided by M. Davachi (personal

communication) DDMI, 1998a) DDMI predicted a blast zone (Figure 3-7) that would

extend into Lac de Gras.

This formula is of the same type that is generated from a scaled distance plot as

shown in Equation 2-9.

( ) 6.1/1727

−= WRPPV (3.1)

Using known blast weight (W) we are able to solve for the distance (R) by setting the

PPV at 13 mm/s. The minimum safe distance based on this very conservative formula is

717.04 m.

Page 64: University of Alberta Library Release Form

47

Figure 3-7) Map of area showing proposed fish incubator sites (black squares). Circular contours represent distance from the centre of A154N/S, contours each 500-m and the red line indicates the DDMI predicted zone where PPVs would exceed DFO guideline levels.

Based on a typical charge weight of 1140 Kg the minimum safe distance to keep the

PPVs at or under the mandated 13 mm/s, calculated using the simplified DFO formula

(equation 2-22), is 509.5 m.

Both formulae predict a zone where PPVs will exceed 13 mm/s will extend beyond

the dike wall and into the lake substrate, and therefore, into fish habitat:

DFO; 510 metres from the blast centre or up to 200 metres into the lake.

DDMI; 717 metres from the blast centre or up to 407 metres into the lake.

Page 65: University of Alberta Library Release Form

48

3.4.2) Over Pressure

The over pressure, or Instantaneous Pressure Change in the lake waters due to

blasting on land can be calculated based on DFO formulae discussed in chapter 2. Firstly

we predict the peak particle velocity (eqn. 3-2). Assuming a minimum distance of 310 m

to the outer edge of the dike and an average charge weight of 1140 kg we calculate a

maximum particle velocity of 4.97 cm/s at the edge of the substrate. The pressure in the

substrate is then calculated by solving for Pr (kPa) from equation 2-6.

2/* rrr CDPPVP = (3.2)

PPV is the predicted Peak Particle Velocity (mm/s) and Dr is the density of the

substrate (2.04 kg/m3) and Cr is the compressional wave velocity in saturated soil (146.3

m/s). Then from equation 2-4 the pressure transferred into the water column is found to

be 74.2 kPa.

Using DDMI guideline calculations we can determine the minimum offset needed to

limit the pressure transferred into the substrate to 100 kPa. Assuming a charge weight of

1140 kg, a substrate velocity of 5,000 – 6,000 m/s and a water velocity of 1,453 m/s we

can calculate the minimum safe distance.

Knowing the densities of the substrate and the water we can use equation 2-4 to

calculate that, the amount of pressure that gets transferred from the substrate into the

water column is only 14%. A pressure of 714 kPa in the substrate would transfer 100 kPa

into the water column. To achieve this pressure a particle velocity of between 116.7 and

142.8 mm/s (depending on the substrate velocity) is needed (from equation 2-6). Based

on the PPV we can calculate the distance using equation 2 – 8 to solve for R. The

minimum safe distance for keeping the induced over pressure transferred into the water

Page 66: University of Alberta Library Release Form

49

column by an 1140 kg blast is 113.9 m for a substrate velocity of 5,000 m/s to 129.3 m

for a substrate velocity of 6,000 m/s.

Another calculation can be made to address the transfer of seismic energy across an

interface. Two equations for transfer and reflection of energy (Telford et al 1990) are

shown. The transmitted (Et) and reflection (Er) of energy for a P-wave normally incident

on the interface between the rock and the water, initially propagating in the rock

substrate, and transmitting as a P-wave into the water are given by:

2

12

12

+−

=ZZZZ

EE

i

r (3.3)

( )2

12

214ZZZZ

EE

i

t

+= (3.4)

where the acoustic impedance Z is the product of the material density and compressional

wave (P-wave) velocity and Ei is the original energy of the wave. Assuming the same

values for density and P-wave velocities as above, we can show that less than 40% of the

energy is transferred into the water column. The problem becomes much more

complicated if one accounts for variations in the angle of incidence of the seismic wave

with respect to the rock-water interface (i.e. the lake bottom). Examination of the results

of more detailed calculations (e.g. Stein and Wysession, 2003) show that generally the

maximum transmission occurs at normal incidence and hence the above equations

provide the maximum values of transmission. Of course, usually the angle of incidence

of these body waves will be less as the lake bottom will be at some angle with respect to

the incoming wavefront.

Page 67: University of Alberta Library Release Form

50

As an exercise, it is useful to see some relationships between overpressures in the

water and particle motions in the rock. Eqn. 3-4 could also be written in terms of the

particle motions with

( )12

12ZZ

ZAA it +

= (3.5)

where, At and Ai are the transmitted and incident particle displacements. Examination of

Eqn, 2.2 will show that the amplitudes can just as easily be given in terms of the particle

velocities which are measured. Consider the case with the physical properties for the rock

of Vp = 6000 m/s and ρ = 2700 kg/m3 and with lake water with Vp = 1450 m/s and ρ =

1000 kg/m3 which yields At/Ai = 1.84. For the sake of argument, consider a high

amplitude incident P-wave in the rock with Ai = 10 mm/s; this gives At = 18.4 mm/s

which equates to an overpressure in the water of P = Z2At = 26.6 kPa. Note, that this is

for a body wave and a value of 10 mm/s for the P-wave propagation is a very strong

pulse. As noted earlier, however, the main particle motion of the passing wave is due to

propagation of the Rayleigh waves which provide substantially lower coupling of wave

particle motion to pressure in the water. Consequently, this rather simple analysis

suggests that the overpressures in the water produced by blasting in the hard rock will be

minimal. One should not, however, use these simple estimates to determine overpressures

in the water for a water borne explosion, which is a completely different problem.

3.5) Summary

Using three different approaches we can show the transfer of energy into the water

column is minimal and that the DFO mandated guideline of 100 kPa is not attainable

Page 68: University of Alberta Library Release Form

51

based on typical explosive sizes and distances from the lake. On the other hand several

different approaches show that PPVs in excess of 13 mm/s may potentially reach well

beyond the dike and into the lake substrate. DDMI suggests that harmful PPVs can

extend up to 407 metres into the lake while DFO calculations suggests up to 200 metres

into the lake. As such, DDMI teamed up with the University of Alberta departments of

Biological Sciences and the Institute for Geophysical Research to monitor blasting and

study the blast effects on fish habitat within Lac de Gras

Analysis of the seismic data gives some insight to the seismic properties of the host

rock. The passive data suggests a Rayleigh wave velocity of about 2600 m/s and the

refraction survey indicates a surface wave velocity of about 1000 m/s.

Page 69: University of Alberta Library Release Form

52

Chapter 4) Methodology; Field component

The objectives of the overall field study are to 1) determine if blasting causes

increased mortality among incubating lake trout eggs and 2) to determine if blasting

effects on fish habitat at Lac de Gras change as the pit deepens. The work in this thesis

focuses primarily on the technical aspects of recording and predicting the blast levels to

complement the parallel work of (Faulkner, 2006). In this chapter, I outline the

development of a specialised system that allowed the simultaneous measurement of both

particle velocities and overpressures at the same point in the lake. This development

relied on some recent innovations with sea floor detectors developed for time lapse

seismic monitoring on the ocean bottom.

To quantify the effects due to blasting, Blast Mate® recorders produced by Instantel,

of Ottawa Ontario, are used to record the movement in the ground and also the pressure

changes in the air. Conventionally, each unit is connected to a 3 component geophone

and a microphone. DDMI generally deploys 4 of these units on or within the dike. Using

a special 4 component sea type sensor that replaces the microphone with a hydrophone to

measure water pressure, the U of A team was able to monitor the same blasts as they are

felt within the lake substrate. Three sites within the predicted blast zone and one well

outside were chosen for blast monitoring.

During the first component of the field study, incubators holding fish eggs were

placed alongside of the sensors. Mortality in the fish eggs was counted and compared to

monitored and predicted PPVs.

Page 70: University of Alberta Library Release Form

53

The second component of the field study encompasses all blast monitoring data

collected at DDMI for analysis to determine the changes in blasting effects as the pit

deepens.

4.1) Site Selection

Four sites were selected for sensor placement (Figure 4-1), based on the quality of the

lake bed for fish habitat and the distance from the pit. Informally they are known for their

location and recorder serial number as South Dike (8902), Reference Island (8903), Tern

Island (8904) and East Dike (8905).

Lake trout move onto rocky shallows inshore or on shoals in mid-August prior to

spawning. Spawning habitat suitability is generally based on depth (2.5-7 m but next to

deeper water), substrate size (4-30 cm, but up to 1 m) and shape (angular), interstitial

space (20 cm – 1 m deep), and being located away from depositional effects (Scott and

Crossman 1973, Gunn 1995, Marsden et al. 1995a, DDMI 1997b). These eggs then

incubate over winter and hatch 8-10 months later (May to July in Lac de Gras). It is over

this long incubation period they are exposed to the ambient environmental conditions,

and to disturbances such as blasting.

The first three sites are all situated near the dike and within the DDMI predicted blast

zone while the Reference Island site is selected to provide baseline measurements (Figure

4-1.). The reference site is chosen for its distance from the blasting operations where 3

km is considered to be far enough away that shaking and PPV values are negligible. The

East and South dike sites were selected at strategic locations along the dike to allow for

ease of access. A fourth site at Reference Island, about 1 km offset, was selected to

Page 71: University of Alberta Library Release Form

54

provide a baseline measurement where it is believed that blasting will have none, or

negligible, effects.

Figure 4-1) Locations of the 4 sensors in relation to the dike, proposed pit and predicted Blast Zone (Red Line). Distance between scale marks is 1000 m of the UTM grid. DDMI sensors are deployed at various locations along the dike. Note that the Blast Zone takes into account the location of the next pit south and east of the original.

The East Dike and South Dike locations are easily accessible directly on the dike

while the 2 other sensors are accessible only by boat during the summer and by

snowmobile when the lake is frozen. During winter freeze-up and spring thaw or during

extreme weather conditions these two sensors cannot be safely reached. As such, the data

set from the latter sensors are not as complete as that of the other two.

Page 72: University of Alberta Library Release Form

55

4.2) Sensor Placement

Sensors were placed in the four selected sites on September 9th of 2003. All of the

sensors were placed at a depth of 3 to 5 metres together with the fish incubators. This

depth was chosen in order to avoid as much as possible the lake ice during winter.

Detailed descriptions of these incubators may be found in Faulkner et al. (2006) but,

briefly, these consisted of 50 cylindrical ‘cells’ produced by drilling holes through a 12.5

cm x 25 cm x 1 cm Plexiglas sheet.

Eggs were placed into each of the cells and then sealed using a fine mesh that was

intended to keep the developing embryos safe from predation and escape. Twenty

incubators were placed at each site. Ten incubators were collected from each site after 20

days, related to the anticipated fish development.

Divers secured the chains, the incubators and the sensors to the boulder strewn lake

bottom. The sensors were attached to a chain along with the incubators and also wedged

under rocks to ensure coupling with the substrate (Figure 4-2). The sensors were then

connected to the on-land recording devices by a 100 m. cable allowing for access from

dry land. On land, the recording units were protected by a standard insulated cooler and

kept warm using rice filled bags that were heated prior to deployment.

Page 73: University of Alberta Library Release Form

56

Figure 4-2) Placement of the 4-C sensor. The sensor is attached to the chain and also placed under rocks to ensure coupling.

4.3) Equipment

An OYO Sea Array 4™ 4-component underwater geophone was used to measure the

ground vibrations and over pressure, attached to an Instantel Minimate Plus™ recording

unit. Data processing was done using the BlastWare Series III Software (Release 4.37)

and other standard software packages.

The 4-components on this specialised underwater package consist of 3 orthogonal 10

Hz. geophones that provide the PPV measurement of the motion of the solid substrate

and a single 10 Hz hydrophone that provides a measure of the pressure within the

overlying water column (Figure 4-3). This package is only recently commercially

available as such detectors are now used on a routine basis for deep water studies in the

North Sea where the combination of the 4 different components adds greatly to data

Page 74: University of Alberta Library Release Form

57

quality. The 3 orthogonally oriented geophones are mounted on gimbals that ensure that

two are always horizontal and the third is vertical. Availability of this unit greatly

simplified the present field tests. The detectors were matched and calibrated with the

recording units by personnel at Instantel prior to field deployment.

Figure 4-3) A cutaway view of the OYO 4 Component sensor. The gimballed geophones can be seen and the hydrophone is situated at the left end of the sensor.

The Minimate sensors acquired for this project had an expanded memory capacity

allowing for 1200, 1 second records at a sampling rate of 1024 samples per second.

Sampling rates available to us were 1024, 2048 and 4096 samples per second. The units

continuously record leaving a .25 second buffer in the memory. When a preset trigger

level is achieved, the unit will record for a desired amount of time plus up to .25 second

pre-trigger waveform.

4.4) Data Collection

Prior to each blast, whenever outside conditions and safety considerations allowed,

DDMI personnel would connect the blast mates to the sensors to record both the PPV and

the overpressure due to the blast wave. Travel to two of the four study sites was

hampered by extreme weather conditions, freeze up, and spring thaw, resulting in

Page 75: University of Alberta Library Release Form

58

incomplete data collection for these sites. During the period of September 9, 2003 and

July 27, 2004, 97 blasts were conducted but of the possible 388 records only 194 were

recorded. The remaining events were not recorded due to 1) extreme weather conditions

when travel to sites was not possible, 2) the recording unit was not properly connected or

3) the blast was not large enough to trigger (i.e. commence recording) by the recording

units.

Three of the sensors are set to trigger on either of the geophone channels at 1.5 mm/s

while the Reference Island site sensor (see Figure 4-1.) is set at the minimum threshold of

.51 mm/s. Often the units were left out for longer periods up to 2 days and recorded many

non blast related events due to the low trigger level. The units continuously record and

hence are able to store a ‘buffer’ of pre-trigger time data for up to .25 seconds. The

memory on the units allows for a continuous record for only 1 to 13 seconds. Technical

specifications are provided in Appendix 1. All units were set to pre-trigger by 0.2 seconds

and to record for 10 seconds. Generally, the blast induced vibration lasted less than 5

seconds. The extended memory capacity of the unit allowed for recording of up to 30 of

these 10 second events. The highest possible sampling rate of 4096 samples per second

was chosen; this corresponds to a Nyquist (i.e. peak resolvable frequency) of about 2 kHz

which was well above the frequency content of the blast signals as shown in Chapter 3.

After blasting the recording units were collected and the data were downloaded to a

PC and sent to the U of A for analysis. The data were routinely scrutinized by University

of Alberta personnel to determine if recorded events can be attributed to the actual

blasting (Fig. 4-4) or due to mistriggers from non-blast related events (Figure 4-5) or

electrical noise (Figure 4-6). The desired blast related events display a very distinctive

Page 76: University of Alberta Library Release Form

59

waveform from other non-events. A further difference was that the non-blast events

would rarely trigger all of the instruments. A time correlation between blast and record

time removed any final doubt as to the source of the signal. Each record includes Peak

Particle Velocity for each channel (geophone and hydrophone) and its corresponding

frequency, the time of the peak in relation to the trigger, the peak acceleration and the

peak displacement. A plot of all of the waveforms for the 10 second recording period is

presented along with a plot of Velocity vs. Frequency. Also the largest peak vector sum is

calculated over the entire event. A typical blast record produced by the Instantel software

is presented in Fig. 4-4, however all the original digital data were obtained to allow for

additional analyses.

Figure 4-4) Typical waveform due to a blast as presented using the Instantel Blast WareTM software. The channels are overpressure (top), and the 3 geophone channels, longitudinal, vertical and transverse. The vertical scale is 10 mm/s for the PPV and 10 pa for the overpressure per division. Inset) Frequency spectrum for each of the geophone channels; Y-axis Particle Velocity (mm/s) and, X-axis Frequency in Hz. Both Axes are log scale.

Page 77: University of Alberta Library Release Form

60

Figure 4-5) Blast record for a non blasting related event. As above, the channels are overpressure (top), and the 3 geophone channels, longitudinal, vertical and transverse. The vertical scale is 2 mm/s for the PPV and 10 pa for the overpressure per division.

Figure 4-6) A typical waveform due to electrical noise. As above, the channels are overpressure (top), and the 3 geophone channels, longitudinal, vertical and transverse. The vertical scale is 10 mm/s for the PPV and 10 pa for the overpressure per division.

Page 78: University of Alberta Library Release Form

61

4.5) Data Processing and analysis

To allow for analysis of the data, collected as the Amplitude of the Particle Velocity

over Time, the corresponding blast data (i.e. blast weight per delay and location) is

needed. Knowing the scaled distance and the PPV standard blast monitoring techniques

are used to predict safe blasting levels. Scaled distances were calculated for all available

U of A data. Logarithmic plots were created for each site using the PPV in mm/s as

provided directly by the Instantel Blast WareTM software and as checked independently

by our own calculations.

The data can be further analysed as in the second part of the study by comparing

changes in the components of the PPV prediction function.

Page 79: University of Alberta Library Release Form

62

Chapter 5) Field Component Results

The objective of the field study was to determine what, if any, were the effects on the

incubating fish eggs and to relate these effects to mortality among the eggs. To achieve

this, Faulkner et al. (2006) assessed the mortality in fish eggs to relate it to the vibration

exposure at each site. The blast data was then analysed to produce a PPV prediction

function for each site and to redefine the blast zone.

The second objective of the field component was to monitor changes in blasting

effects as the pit deepens and to assess if geophysical and biological effects of blasting

attenuate over time as the pit deepens.

5.1) Analysis

During the period of September 9, 2003 through July 27 2004, 107 blasts were (Data

is presented in Appendix 4) conducted; of the total 428 possible recordings (107 for each

of the four sites) only 184 were recorded. Of these 184 events 8 were found to have

exceeded the Department of Fisheries and Oceans (DFO) guideline of 13 mm/s Peak

Particle Velocity (PPV) at three of the four study sites. The largest recorded event was

28.5 mm/s. The prediction function suggests that up to 72 blasts during this time may

have exceeded guidelines and reached up to a maximum PPV of 28.8 mm/s.

To complete their study, Faulkner et al. (2006) required a PPV value for each blast at

each incubator in order that appropriate vibration and overpressure exposure levels could

be determined. As there were gaps in the data set, shaking levels for each blast that was

not recorded were predicted using one of two empirical prediction functions. One method

involved utilising the prediction function derived by DDMI using their earlier collected

Page 80: University of Alberta Library Release Form

63

data and prediction function. The second involved analysis and creation of our own

empirical prediction function from the observations covering the period while incubators

were in the lake.

Transmission of the blast energy into the water column from the rock substrate was

not consistently measured, however, it is considered to be of little effect, see Chapter 3,

and in fact the largest measured pressure was only 11 Pa, a value that was barely

detectable, and as such only results from ground vibration are discussed in the remainder

of the thesis as overpressure does not appear to be of concern.

5.2) DDMI Prediction Function

The DDMI blast monitoring program has developed a scaled distance plot from their

collected data at the site. These data likely were collected early in the blasting

programme as they do not fit analysis done by us.

)/(700 6.1 smmSDPPV −×= (5-1)

The upper and lower bounds are defined as 2 x and 0.5 x the PPV. It is interesting to note

that the maximum predicted PPV scalar multiplier is 1400 (i.e. max PPV = 1400xSD-1.6)

while the initial prediction (Chapter 3.) used a value of 1700.

For our own analysis of the data collected by DDMI from August 27, 2002 to January

17, 2004, we find, through regression analysis, the following function:

)/(510 5.1 smmSDPPV −×= (5-2)

Using the same methodology as that at DDMI and Ekati, the lower and upper bounds are

2 x and 0.5 x the PPV.

Page 81: University of Alberta Library Release Form

64

Figure 5-1) The DDMI prediction function (Equation 5-2) presented in red lines with the data collected by DDMI from August 27, 2002 to January 17, 2004. The black lines represent the prediction function derived from the plotted data (Equation 5-2).

The DDMI prediction function (Figure 5-1) has a good fit with the data but it can be

seen that many more events fall below the lower bound than fall above the upper bound.

To best use these functions to predict PPVs, it is important to place sensors within the

area in question. The DDMI sensors are typically placed in 3 different types of locations;

the dike crest, the toe berm and also within the dike which are expected to have a

different elastic behaviour from the original solid rock substrate.

Figure 5-2) Scaled Distance plot of the four U of A sensors compared to the DDMI data. The red lines represent the minimum, median and maximum PPV for a given scaled distance from the DDMI prediction function. 8902, 8903, 8904 & 8905 are South Dike, Reference Island, Tern Island and East Dike respectively.

Page 82: University of Alberta Library Release Form

65

The U of A sensors, are instead, placed within the waters of Lac de Gras on lake

substrate and a better PPV prediction function can be had from data collected there. Data

plotted from our four sensors and compared to the DDMI prediction function (Figure 5-2)

show poor correlation; they are not well contained within the bounds and display a

different slope.

Consequently a linear regression analysis was done on our data and the best fit line

and confidence limits were found for each of the four sensors’ data sets and then the PPV

was predicted for each blast at each location. Finally a blast zone where the PPV may

exceed 13 mm/s was derived using data from the 3 sites situated near the dike.

Each of the sensors is presented separately with data from the three sensors combined

to derive an updated blast zone. As a result a more useful predictive function can be

established enabling Faulkner et al to have a range of PPV values that can be used for

their analysis.

5.3) South Dike (8902)

The South Dike sensor was located just outside of the dike wall (Figure 4-1). The

distance from the sensor to the minimum, mean and maximum distance to the blast

locations is 335 m, 744 m and 1059 m respectively. Of the 78 events recorded at the

South Dike location, 4 exceeded 13 mm/s. Using the prediction function (5-3) no events

are predicted to exceed the guideline while using the upper bound, up to 28 events could

exceed 13 mm/s.

PPV = (SD-0.732) x 53 (mm/s) (5-3)

Page 83: University of Alberta Library Release Form

66

5.4) Reference Island (8903)

The Reference Island sensor was located well outside of the dike wall (Figure 4-1) at

an average distance of 3014 m. from blast location. Minimum and maximum distances

are 2570 m and 3350 m respectively. Of the 10 events recorded at the Reference Island

location, none exceeded the guidelines and in fact, most events did not even trigger the

recording unit. Using the prediction function (5-4), no events can exceed the guidelines.

Based on the regression analysis we can predict the PPVs experienced for all blasts.

PPV = (SD-0.732) x 1.3 (mm/s) (5-4)

5.5) Tern Island (8904)

The Tern Island sensor was located on an island situated about 100 m away from the

dike wall (Figure 4-1) at an average distance of 940 m from blast location. Minimum and

maximum distances are 540 m and 1240 m respectively. Of the 31 events recorded at the

tern Island site, 2 exceeded 13 mm/s. Of the total 107 events it is predicted that none

exceed guidelines (5-5) using the best fit line while up to 24 events could exceed 13

mm/s using the upper bound to predict the PPVs.

PPV = (SD-0.81) x 88 (mm/s) (5-5)

5.6) East Dike (8905)

The East Dike sensor was located just outside of the dike wall (Figure 4-1) at an

average distance of 875 m from blast location. Minimum and maximum distances are 425

m and 1185 m respectively. Of the 75 events recorded at the East Dike location, 2

exceeded 13 mm/s including the highest recorded value of 28.5. Of the total 107 events it

Page 84: University of Alberta Library Release Form

67

is estimated that none exceeded guidelines using the best fit line while up to 20 events

could exceed 13 mm/s using the upper bound to predict the PPVs.

Based on the regression analysis we can predict the PPVs experienced for all blasts.

PPV = (SD-0.75) x 65 (mm/s) (5-6)

The data is presented here graphically, firstly we have the four regression functions and

then we show the plot of predicted PPVs for each site.

Figure 5-3) Logarithmic plot of the measured particle velocity and scaled distance for the South Dike location. The lines indicate the minimum, median and maximum PPVs to be expected at the site (r-square .26).

Figure 5-4) Logarithmic plot of the measured particle velocity and scaled distance for the Reference Island location. The lines indicate the minimum, median and maximum PPVs to be expected at the site(r-square .09).

Page 85: University of Alberta Library Release Form

68

Figure 5-5) Logarithmic plot of the measured particle velocity and scaled distance for the Tern Island location. The lines indicate the minimum, median and maximum PPVs to be expected at the site(r-square .27).

Figure 5-6) Logarithmic plot of the measured particle velocity and scaled distance for the East Dike location. The lines indicate the minimum, median and maximum PPVs to be expected at the site(r-square .26).

Page 86: University of Alberta Library Release Form

69

Figure 5-7) Minimum, maximum and median predicted PPVs plotted along with actual recorded values for South Dike. The line indicates the DFO guideline value of 13 mm/s.

Figure 5-8) Minimum, maximum and median predicted PPVs plotted along with actual recorded values for Reference Island. Note that the vertical scale is much smaller in this case.

Figure 5-9) Minimum, maximum and median predicted PPVs plotted along with actual recorded values for Tern Island. The line indicates the DFO guideline value of 13 mm/s.

Page 87: University of Alberta Library Release Form

70

Figure 5-10) Minimum, maximum and median predicted PPVs plotted along with actual recorded values for East Dike. The line indicates the DFO guideline value of 13 mm/s.

5.7) Blast Zone

DDMI defined their blast zone and suggested that it could reach as much as 407 m

beyond the dike wall while the DFO calculations suggest that it could reach as far as 200

m. Using the DDMI data and the charge weight of 1140 kg, using our regression

methodology (Eq. 5-2) minimum safe distance would be 619 m from the blast (309 m

into the lake). The best measure, however, of blasting effects in the substrate is to analyse

data that were collected there. We can combine the data from the 3 sensors that were

situated near the dike wall. The East dike, south dike and tern island data sets are exposed

to similar PPV readings and are situated at similar distances from the blasting area..

Based on the regression analysis (Figure 5-11), of the 3 data sets, we can predict the

PPVs for all blasts.

Page 88: University of Alberta Library Release Form

71

Figure 5-11) Logarithmic plot of the measured particle velocity and scaled distance for the East Dike, South Dike and Reference Island combined. The lines indicate the minimum, median and maximum PPVs to be expected (r-square .24). .

PPV = (SD-0.72) x 56 (mm/s) (5-7)

Once again, assuming a blast weight of 1140 kg, the minimum safe distance would be

672 m meaning that 13 mm/s could be exceeded up to 362 m into the lake.

Blast weights used at DDMI, however, range from 71 kg to 3000 kg with mean and

median values at 1012 kg and 985 kg respectively. Assuming maximum predicted PPV

and 3000 kg weight, 13 mm/s can be attained in the ground for as far 1090 m (780 m

beyond the dike wall).

Distances are calculated for a range of charge weights (Figure 5-12) from equation 5-

7 to predict both the median and maximum distances before PPVs drop to the DFO

guideline of 13 mm/s.

Page 89: University of Alberta Library Release Form

72

Figure 5-12) PPV predictions for different charge weights and distances based on data from the 3 near dike sites. The dashed line represents 13 mm/s. The Predicted values are based on the upper bound of the function.

We are able to predict PPVs for events that were not recorded to complete the data

needed for the study by Faulkner et al. This is done from data collected by the U of A

team. This data set is then used to redefine the blast zone where PPVs may exceed 13

mm/s based on charge weights used in the mining operations.

5.8) Changes in effects as the pit deepens

From the beginning of the mining operation until June 2005 the open pit has

deepened by 90 metres. Analysis of the blast data collected by DDMI and of our own

data is done to assess if geophysical and biological effects of blasting attenuate over time

as the pit deepens.

Two methods were employed to track changes: a) a normalised PPV is calculated by

normalising the Scaled Distance to see trends in the measured PPV over time, and, b).

Page 90: University of Alberta Library Release Form

73

Scaled Distance plots are calculated for 2 month intervals to see the changes in the

prediction function variables A and M.

Since the 27th of August, 2002 when production and blasting began at DDMI 1243

blasts were recorded through June 29, 2005. 1116 of these blasts have grid co-ordinates

and, hence, elevations. The elevations (Figure 5-13) range between 328 and 420 metres.

The elevations of the blasts tend to decrease as the pit deepens, but as is seen, some

blasting is still being done at higher benches. The figure also shows the tendency to blast

on benches at every 10 metres.

Figure 5-13) Blast elevations at DDMI since January 2003. Over time the elevations have decreased as the open pit operation increases in depth.

Analysis of the recorded PPVs over time (Figure 5-14) shows that, while extremely

high PPVs are not frequent there are extremes in both May 2003, and March 2004.

Further analysis shows that these data points are due to blasts being very close to sensors,

(100 m. or less). A normalisation for the scaled distance should remove the extremes.

Nonetheless, no trend is apparent.

Page 91: University of Alberta Library Release Form

74

Figure 5-14) PPVs (mm/s) as measured by DDMI sensors since January 2003.

5.8.1) Changes in normalised PPV

From the recorded blasts, the minimum, maximum and mean values of the scaled

distances are 1.9, 74.2 and 11.9 respectively. The mean value of 11.9 is chosen to be used

as a normalisation factor to calculate a normalised PPV

9.11

SDPPVPPVn ∗= (5-8)

where PPVn is the normalised peak particle velocity and SD is the scaled distance. The

normalised PPV values can be compared against date, elevation, distance and the blast

weight.

Once the normalised value is plotted against date no discernible trend is seen (Figure

5-15). Comparing PPVn to the blast elevation also shows no trend (Figure 5-16).

Comparing PPVn to the blast weight or to the distance is trickier as they are both

components of the scaled distance which is the normalisation factor. As distance to the

blast increases, PPVn decreases exponentially, in a similar manner to that of the Scaled

Distance plot (Figure 5-17).

Page 92: University of Alberta Library Release Form

75

Figure 5-15) Normalised (DDMI) PPVs plotted against the date.

Figure 5-16) Normalised (DDMI) PPVs plotted against the blast elevation.

Figure 5-17) Normalised (DDMI) PPVs plotted against distance from blast. As the distance increases, the measured PPV decreases.

Page 93: University of Alberta Library Release Form

76

Figure 5-18) Normalised (DDMI) PPVs plotted against the blast weight.

Comparing PPVn and the blast weight, no real trend is seen (Figure 5-18). The

normalisation factor accounts for changes in the blast weight.

5.8.2) Changes in the Scaled Distance function

An alternative analysis of changes over time is to see how the Scaled Distance

function evolves over time and with change in pit depth and distance from the blast. For

the DDMI data, Scaled Distance functions are created for small time, depth and distance

intervals for comparison.

In the empirical PPV prediction function (Eqn. 2-29) A & M are the scalar and

exponential variables. These variables are analysed and presented (Figures 5-19 to 5-21).

Page 94: University of Alberta Library Release Form

77

Figure 5-19) Change in A & M over time. (DDMI data). The scalar variable, A, is on the left axis and M, the exponential variable is on the right axis.

Figure 5-20) Change in A & M as the depth increases (elevation decreases). (DDMI data). The scalar variable, A, is on the left axis and M, the exponential variable is on the right axis.

Figure 5-21) Change in A & M as the distance from the blast increases. (DDMI data). The scalar variable, A, is on the left axis and M, the exponential variable is on the right axis.

Page 95: University of Alberta Library Release Form

78

It is interesting to note that the trends of the two variables are opposite suggesting that

they balance each other out. As is earlier noted, changes over time are not a good

measure of changes with increase in depth. In fact, over time neither A, nor M change

much (Figure 5-19). As the depth of the pit increases, A decreases and M increases

(Figure 5-20). The opposite is found as the distance from the blast increases (Figure 5-

21). The question here is; are the changes in A balancing out the changes in M within

error bars or are there actual changes in the SD functions as either distance, or depth into

pit, change?

Page 96: University of Alberta Library Release Form

79

Chapter 6) Laboratory Study

Although the maximum allowable PPV limit for protecting spawning beds is a precise

albeit somewhat arbitrary number, previous attempts to relate physical effects of blasting

to mortality of incubating eggs has failed to show any increases in mortality. Our work

(Faulkner et al. (in press), Faulkner thesis) found no evidence for elevated mortality of

lake trout eggs from the blasting exposures at DDMI’s operations at Lac de Gras as

described in detail in the previous chapter, despite exposures more than double the PPV

guideline (28.5 mm/s). However, the field setting did not allow for proper control

regarding the number, timing, and sizes of blasts to allow researchers to properly

determine the critical PPV levels at which survival of fish eggs decrease. The objective of

the laboratory study was to emulate field blasting conditions and their effects on fish and

to find a PPV level where an increased mortality is seen.

A laboratory experiment was designed where incubating Rainbow Trout eggs could

be exposed to particle motions as great as 250 mm/s. Blast events were simulated in a

tank and measured by identical equipment to that described in chapter 3. PPV,

acceleration, frequency and duration of blast event are all criteria that were noted and

improved whenever possible to better simulate Diavik-like conditions. To allow for direct

comparison of blasting effects between lab and field events, a normalised ‘pseudo

energy’ waveform attribute was calculated.

Three sets of laboratory experiments were conducted and as is common in

experimental developments, each consecutive experiment was an improvement on the

previous. New technical developments allowed the final round in September of 2005 to

Page 97: University of Alberta Library Release Form

80

be much more repeatable with a small error on the PPV reduced to only ~1.5%. This

allowed the final experiment to be conducted without the sensor in the tank thus allowing

for more room in the tank. In fact, the eggs were placed in a gravel bed to better simulate

the field conditions. Only the results from the third, and final, study were used (Faulkner

et al. 2006) and only these results will be reported here. In this interdisciplinary effort,

Faulkner et al. (2006) was primarily responsible for detailing the statistical results of the

vibration exposure tests while this report provides the details of the experiments.

Consequently, while the predominant part of this study was biological in nature, the

development and calibration of the instrumentation and the analyses of the blast and

laboratory data were carried out primarily by the author.

6.1) Methodology

A weight drop experiment was designed whereby fish eggs were exposed to a

controlled vibration exposure that partially simulated a blasting event. The experimental

apparatus consisted of a 2 m X 1 m X 0.635 cm metal plate used as the supporting

substrate. A 2 m high pole (Figure 7-1) attached to this plate guided dropped weights

released from known heights h. Ignoring the effects of friction of the weight and of air

resistance, the potential energy Ep provided by the dropped weight with mass M was

linearly proportional to the height h according to Ep = Mgh. The impact of the weight on

the plate sent a simulated blast wave through the plate and the small tank where the fish

eggs were placed.

To measure and record the effects of the simulated blast, the same Oyo 4-C sensor

used in the field tests was placed inside the tank and its responses recorded during each

tests. As with the field experiment, the Instantel Minimate Plus™ was automatically

Page 98: University of Alberta Library Release Form

81

triggered by the event. Five second records were kept with a pre-trigger of 0.25 seconds.

Work was done to ensure that the 5 heights used would repeatedly produce a known a

shaking measured in PPVs. To allow for more room in the tank, the sensor was removed

for the experiment and the calibrated heights were used to reproduce a given PPV.

Figure 6-1) The weight-drop apparatus; a: steel pole, secured at top with straps (not shown); b: pin to release the weight apparatus; c: the weight set-up, including an aluminium sleeve and cast iron weights; d: holes drilled in the pole spaced 15 cm apart; e: 0.64 cm thick rubber mat; f: 0.64 cm thick steel base plate; g: 5 cm foam padding; h: 2.5 cm rubber mat, and i: fibreglass tank.

6.2) Results

For the third and final component of the lab study, five blasting levels were used.

Figures 6-2 and 6-3 show the record for two of the levels that were used, 138 mm/s and

245 mm/s. Each level was repeated 10 times prior to actual exposures to confirm a PPV

and the Standard Deviation. Note that the vertical scale is in mm/s and is set the same for

each different height to show the differences in intensities.

Page 99: University of Alberta Library Release Form

82

0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2

-200

-100

0

100

200

Tran

s

Particle Velocities vs. Time for all 3 components

0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2

-200

-100

0

100

200

Ver

t

0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2

-200

-100

0

100

200

Time (s)

Long

Figure 6-2) PPV = 138 mm/s Height = Standard Deviation (2.4)

0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2

-200

-100

0

100

200

Tran

s

Particle Velocities vs. Time for all 3 components

0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2

-200

-100

0

100

200

Ver

t

0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2

-200

-100

0

100

200

Time (s)

Long

Figure 6-3) PPV = 245 mm/s Height = Standard Deviation (3.0)

Page 100: University of Alberta Library Release Form

83

6.3) Pseudo energy

Although the PPV standards have been used for a long time and are well accepted in

various standards as noted in previous chapters, it must be remembered that the PPV

measurement is only one simple attribute of the shaking. Indeed, one could really ask

whether or not the PPV measure itself, while readily measured and easily understood,

really tells us that much about the propensity of a blast to cause damage. Indeed,

Dowding (1985) in his classic text makes the case that workers should not ignore the

frequency of the blast vibrations. This is particularly germane to his application of

building damage as most buildings will have a particular set of resonant ‘modes’ at given

frequencies for which particle motions are amplified and can cause particular damage. It

is these modal vibrations in buildings that earthquake engineers strive to dampen. The

PPV value really is only one single measure or attribute extracted from the entire

waveform. It contains no information about other factors that may be more important,

such as the frequency content, the duration of vibration, or the total energy of vibration.

While we understand the reason why field engineers prefer the simple PPV measure, we

find it surprising that these other factors that may relate more to an overall vibration

‘exposure’ which might be more important to predicting vibration damage levels have

been ignored. An implication of focusing on simpler measures might be that the legal,

social, or professional criteria assigned could be overly conservative and more carefully

designed and scientific assessments might even allow for greater blasting levels.

We do not expect to solve this problem as part of this study. However, it is interesting

to begin the search for more informative waveform attributes. To allow for a truer

comparison between blasting events an estimate for the energy, called the pseudo-energy

Page 101: University of Alberta Library Release Form

84

Es, was developed. Es takes into account the entire waveform, including its duration and

all the particle velocities, via:

[ ]∑=tn

ts tPVE

0

2)( (6-1)

This stems from the formula for the instantaneous kinetic energy KE(t):

[ ]∑

=

tn

ttPVmtKE

0

2)(21)( (6-2)

where m = mass, PV = particle velocity (not wave propagation velocity), and t0 and tn

define the beginning and end times of the waveform of duration tn – t0. Because in this

application mass is constant, kinetic energy is directly proportional to PV2. Therefore, to

compare between events, this total energy was calculated. The squared particle velocities

of the three geophone channels were summed and plotted over time, and the area under

the curve was determined. By calculating the area under the curve, we have invariably

taken into account the frequency and duration of the event. A longer event transmits and

produces more energy and this is reflected in the calculation. However, it should be noted

that this calculation allows for comparison between different events and is in no way a

true energy calculation.

Figure 6-4 shows the 3 component waveform forms for a lab simulated event and

an event recorded at Diavik. For each event the calculated pseudo energy is shown.

Page 102: University of Alberta Library Release Form

85

Figure 6-4a)

Figure 6-4) Blast records for a lab simulation event (a) and DDMI blasting event (b). The bottom panel shows the 3 component wave form and the top panel shows the pseudo energy calculated from the vector sum of the waveform. Note that the amplitude scales are the same for comparison sake but the time scales are different. The DDMI event is much longer than the lab event.

Page 103: University of Alberta Library Release Form

86

Chapter 7) Discussion

7.1) Field component

DFO guidelines suggest that overpressure limits within bodies of water should be

limited to ± 100 kPa (about 1 atmosphere of pressure). The maximum blast water

overpressure observed was 11 Pa. As noted, in most cases the water overpressure could

not be reliably measured due to low intensities. This situation is not surprising and was

expected at the early stages of the study. The reason for this is that the elastic impedance

(the product of the materials sound velocity and mass density) for water (~ 1.5 MRayl)1

differs substantially from that for the rock substrate (~12.5 MRayl). The contrast in these

properties indicates the level of energy that can be transmitted from one medium to the

next, the greater the contrast the less the transfer of energy. In this particular case the

large contrast between the water and the rock substrate would suggest little energy is

transferred into the water column to be detected as a pressure. Indeed the DFO relation

suggests that about 20% of the pressure in the substrate will be introduced into the water

column. A maximum pressure of 14.25 kPa, which is calculated based on a maximum

PPV of 28.5 mm/s, in the substrate could introduce up to 2.85 kPa of pressure into the

water column which is well below the DFO guideline. In fact of more concern is the Peak

Particle Velocity to which the fish eggs may be exposed to.

1 The Rayl is a measure of elastic impedance, 1 Rayl = (kg/m3)(m/s), 1 MRayl = 106 Rayl. Some care must be exercised in the use of this unit as it also exists in the cgs system as (gm/cm3)(cm/s). As this thesis employs only the MKS (Systeme International d’unites) measures, the first definition is automatically assumed. Again, the Rayl is named after our friend Lord Rayleigh.

Page 104: University of Alberta Library Release Form

87

Initially it was expected that the scaled distance function used at DDMI would fit the

data at the 4 fish incubator sites. Instead, the functions for the sites display different

characteristics. The 3 study sites that are closest to mining operations show a slope

(exponential constant) of about -0.75 while DDMI’s slope is about double, or -1.6. At

reference island the slope is -0.17. The steepness of the slope is and indication of how

rapidly the wave intensity decays with distance; as noted earlier this consists of a

geometric component due to wavefield spreading and a material dependent wave

absorption component. A third consideration depends on the stiffness, rigidity, and

density of the rock substrate; the more compliant and less dense a material is the greater

the expected particle motions. The decay due to simple wavefield geometric spreading is

closely proportional to r1/2 for the surface waves that cause the greatest motions here.

The other factors are not so easily determined; indeed, the seismic characteristics of a

given site complicate the analysis of even the highest quality seismometer data and high

quality stations undergo rigorous site characterizations. Together, a steeper slope

indicates higher attenuation of the blast wave as it propagates. Several possibilities exist

for the difference; Dowding (1985) suggests that differences in particle velocity at a

given scaled distance can be attributed to: 1) changes in geological conditions: 2)

different explosive types: 3) variations in the geometry of blast patterns, and 4) errors in

blast timing. He also warns that large differences in charge weight or distance will also

contribute to differences in PPV. Other possible explanations for this difference include:

• Differences in sensors: The fact that the U of A study sensors are not the same as

those employed by DDMI and may be calibrated slightly differently could

contribute. However, this is not thought to be a major issue as the same basic

Page 105: University of Alberta Library Release Form

88

OYO geophone is used in both sensor types and all the instruments had been

calibrated by Instantel. A calibration difference would be a scalar value while the

U of A data shows both scalar and, more importantly, exponential differences.

• Placement of the sensors: The study sensors are all placed in water and well

coupled to the hard rock. U of A sensors are located on solid granitoid substrate.

In contrast, the DDMI sensors are deployed on the softer dike. The dike is made

from crush and boulders and is likely to be more attenuative than the substrate to

seismic waves.

• Differences in distances in relation to blast patterns: Another reason for variations

in attenuation between the two data sets may be the differences in the average

distance from the sensors to the centre of the blast pattern. DDMI sensors are on

average placed at 344 m while U of A sensors (except Reference Island) are at

852 m. The blast pattern itself typically covers an area of 2500 m2.

• Differences in distances in relation to attenuation. As seismic waves propagate,

attenuation of the higher frequency components of these waves is much more

rapid. As such, the waveforms observed by the closer sensors may include high

frequency, high amplitude waves that attenuate before reaching the U of A sensor.

• Changes in blast prediction due to variations in scale function with time:

attenuation value (slope of scaled distance plot) for the U of A data. Another

possibility is that the DDMI scaled distance function may have changed over

time. This theory is disproved by comparing data recorded prior to September 9,

2003 to data recorded after this date (figure 7-1)

Page 106: University of Alberta Library Release Form

89

• Differences in surface wave dispersion. The largest energy in these records is

contained in what are referred to as surface waves. Although not exactly the

same, one can think of the surface waves as being somewhat analogous to the

slowly travelling waves that make up the bow in the water following the motion

of a boat. Different components of these waves will travel at different velocities

and hence the wave further from the shot points will be more spread out in time

than it is closer to the blast. This must also result in a lower peak amplitude

although it will extend the duration of shaking.

Our present hypothesis for the differences between the U of A and DDMI data is a

combination of the placement of the sensors and the attenuation of the higher amplitude

waves closer to the centre of the blast. In fact, due to preferential attenuation, the scaled

distance and prediction functions are only effective over small distances in relation to

actual sensor/blast pattern distance. The effects of surface wave dispersion have not to

our knowledge been examined but this would be beyond the scope of the present study.

Page 107: University of Alberta Library Release Form

90

101

100

101

102

DDMI Blast Monitoring: Scaled Distance vs. PPV

Scaled Distance

Mea

sure

d P

PV

Before Sept 03After Sept 03

Figure 7-1) Comparison of scaled distance data over two periods shows that the distribution has stayed the same. The solid line indicates 13 mm/s.

7.2) Changes in the blasting effects

It should be expected that the blasting effects will decrease as the pit deepens. As the

pit deepens the areas of concern, the dike and the nearby fish habitat are a greater

distance from the blast source and as such should experience lower levels of vibration

due to attenuation of the blast wave. The change in the pit depth though, is only 90

metres while the blasting area has a radius of 500 metres. The slight changes in distance

attributed to the increased depth are insufficient to change the vibrations if this is the only

consideration.

Page 108: University of Alberta Library Release Form

91

It has been shown (Duvall, 1972) that the level of vibration does not change for the

number of blast holes as long as there is a millisecond delay between each successive

blast. This argument is correct for measured PPVs but this measure does not take into

account the duration of an event, which would increase with added blast holes.

Page 109: University of Alberta Library Release Form

92

Chapter 8) Conclusions

To satisfy the initial objectives of this study, blasting at DDMI was monitored to

study its effects on fish habitat within Lac de Gras. Data collected by the U of A team

within the lake were used to measure the PPVs that incubating lake trout were exposed

to, predict the PPVs when instruments did not record a blast, and to redefine the ‘blast

zone’. Data were also analysed for changes attributed to differing conditions.

The defined blast zone is an area where PPVs up to 13 mm/s may be experienced

based on many conditions, including the blast weight per delay. A better reference is a

graph of safe distances based on blast weight (Figure 5-12). Over time, as the pit

deepened, no changes in blasting effects could be attributed to changes in depth of the pit,

which increased by only 90 metres. The geology does not change and blasting occurs

over an area of radius 500 m, so the increased depth only slightly changes the distance

from the blast.

Predicting the Peak Particle Velocities due to blasting can be effectively done using

any of the previously discussed prediction functions that are based on data collected at

DDMI. As distance from the blast increases, the rate of attenuation decreases as is seen

when comparing the data collected within the pit, just outside of the pit and far away

from the pit. This is because the higher frequencies attenuate faster. While all of the

functions could do a sufficient job in predicting PPVs, the best function is one that is

based on data that is in a similar region i.e. (similar distances and/or charge weights), as

such, the data provided to the mortality study is based on data collected at the incubator

sites.

Page 110: University of Alberta Library Release Form

93

Blast monitoring programmes are commonly done and recorded data are effectively

used to predict safe blast charge weights for adherence to local guidelines for ground

shaking and pressure in the air or water column. Simple back of the envelope type

calculations can be made to safely predict initial blast weights. The success of the so

called, Scaled Distance function is inherent as it is created statistically from data that it

predicts. Each new data point is added and the next prediction is calculated.

Of concern is that fact that ground motions created by blasting are complex and can’t

be simply measured as the peak particle velocity. It is not known whether it is simply the

maximum particle velocity, or if the duration and frequency of an event that will cause

mortality in a fish egg. We attempted to make up for some of the deficiencies by

calculating a Pseudo Energy of a blasting event. The pseudo energy calculation gives a

measure of the total exposure for a blasting event.

Contributions of this work

There are two major contributions of this thesis. To some degree these contributions

are more of a technical nature that allowed us to extend the range of applicability of blast

monitoring and also to provide better controlled information in the laboratory.

First, this is the first blasting vibration study, to our knowledge, carried out in which

both particle vibrations and overpressure are measured simultaneously at the floor of a

lakebed. This work was greatly facilitated by recent technical developments in deep sea

seismic exploration in the North Sea, where combined gimballed geophone and

hydrophone packages were constructed and are now commercially available. Without

this new equipment, the required lake floor testing would have been much more difficult

and would have required that we construct our own special detectors. This would have

Page 111: University of Alberta Library Release Form

94

greatly increased the costs of the project in terms of time and funding. As noted, this

technical development has opened new applications for the blast monitoring community

and was consequently recognized by Instantel, a manufacturer of blast monitoring

equipment.

Second, a new methodology for producing controlled and calibrated simulated blast

vibrations in the laboratory has been developed. The technical aspects of this equipment

is described in this thesis; the results of application of the method has been discussed in

Faulkner (2006) and Faulkner et al. (2006). Prior to this, the techniques used to subject

fish eggs to vibration were open to question. In this development, the new mechanism

produces waveforms that are more similar to the actual blast vibrations. Further, the

vibration levels were repeatable.

Recommendations for Future Work

As in many projects, this work has opened up additional questions.

First, it is not exactly clear what the actual physical basis of much of the blast

monitoring prediction formula are actually based, and workers apparently simply

followed the lead of the first workers. It would be useful to re-examine this problem in

detail again from first principles, and while this might not be practical for a day to day

application of blast monitoring it would be more comforting to have some understanding

of the actual physics of the problem in order that one could better guide the development

of predictive equations. Unfortunately, the problem is not as simple as we would like.

There are numerous factors that will influence the blast vibrations, the most important of

which are 1) coupling of the actual blast energy to the earth, 2) determination of the

interference of various waves from the blast pattern on the basis of time and on the basis

Page 112: University of Alberta Library Release Form

95

of the blast geometry, and 3) the influence of the local site at which the receiver is placed.

All of these problems also complicate the analysis of earthquake records, the seismic

source of which consists of a complex 3-D failure plane that takes a certain time period to

actually fail. This problem also plagues those who attempt to monitor large bomb blasts

globally in efforts to detect nuclear events in particular. Indeed, one can essentially mask

these large events by careful preparation of the blast site. For example, the coupling

between the detonation and the earth is substantially reduced by carrying out the

experiment within a large underground cavity. One component of future work might be

additional tests of simple blasts under more controlled conditions.

To some degree the blast monitoring equipment could be improved. The current

systems in place rely on taking out a recording box which still must be triggered and still

have a rather limited ability to store records. The technology for recording and data

storage, however, has changed very substantially in the last few years and it should not be

difficult to update the equipment to record continuously. Having the full record, instead

of the snippets obtained once the unit is triggered by large blasts might be useful. As

well, such records might also allow, if additional dynamic range is added, the better

detection of the much weaker body waves.

Page 113: University of Alberta Library Release Form

96

References

Alberta Government, 2005, Exploration Regulations, Alberta Regulation 214/98 of the Forests Act, the, Mines and Minerals Act, the Public Highways Development Act, and the Public Lands Act, http://www.srd.gov.ab.ca/land/m_geo.html, (accessed August 21, 2006)

Birch, W. J.M and M. Pegden. 2000. Improved prediction of ground vibrations from blasting at quarries. Trans. Instn Min. Metall. Sec t. A: Min. technol.), 109.

Blair, D.P., 2004, Charge weight scaling laws and the superposition of blast vibration waves, Fragblast, , 8, 221-239.

Bryan, D. and R. Bonner, 2004, The Diavik Diamond Mine, Lac de Gras, Northwest

Territories, Canada, Proc. 8th Int. Kimberlite Conf. (cd-rom), Vancouver, 61-65, (Accessed http://www.diavik.ca/PDF/Diavik Field Guide 8IKC.pdf August 23, 2006).

Davis, W.J., A.G. Jones, W. Bleeker,; H. Grutter, DEC 2003, Lithosphere development in the Slave craton: a linked crustal and mantle perspective Source: LITHOS, 71 (2-4): 575- 589

DDMI. 1997b. Technical memorandum #12-2, 1997: Shoal habitat survey, Lac de Gras and Lac du Sauvage

DDMI. 1998a. Environmental effects report, Fish and water. DDMI Internal Report. Diavik Fact Book, DDMI, July 2005.http://www.diavik.ca/PDF/Diavik_FactBook.pdf

(Accessed January, 2006)

Dowding, C.H., 1985, Blast Vibration Monitoring and Control, Prentice Hall, USA. Duvall, Wilbur, I., 1972, Chapter 7, Damage and prediction control. EM 1110-2-3800,

US Bureau of Mines. Faulkner, S.G. 2006, Blasting effects on Salmonid Eggs, M.Sc. Thesis, University of

Alberta. Faulkner, S.G. and W.N. Tonn. 2005, Effects of Explosives on Incubating Lake Trout

Eggs in the Canadian Arctic. (Blasting Effects Study), DDMI Internal Report,. Biological Component. 32 pages.

Faulkner, S. G., W. M. Tonn, M. Welz, and D.R. Schmitt. 2006. Effects of Explosives

on Incubating Lake Trout Eggs in the Canadian Arctic. Submitted to North American Journal of Fisheries Management. pp 31.

Page 114: University of Alberta Library Release Form

97

Faulkner, S. G., M. Welz, W. M. Tonn, and D.R. Schmitt. 2006. Effects of Simulated Blasting on Mortality of Rainbow Trout Eggs. To be published.

Goble, K.A., January 1980, Project to determine effects of seismic activity on water wells, Alberta Energy and Natural Resources.

Goble, K.A., June 1980, Project to determine effects of seismic activity on water wells, Phase II Alberta Energy and Natural Resources.

Grech, M. 1998, True Amplitude Processing in VSPs, M.Sc. Thesis, University of Alberta.

Grogan, A. and McAnuff, P., 2005, Canadian vibration and air blast regulations, presented at the Int. Soc. Explosives Eng. Conference.

Gunn, J.M. 1995. Spawning behaviour of lake trout: effects on colonization ability.

Journal of Great Lakes Research 21(S1): 323-329. Lai, C. G., G. J. Rix, et al. 2002. "Simultaneous measurement and inversion of surface

wave dispersion and attenuation curves." Soil Dynamics and Earthquake Engineering 22(9-12): 923-930.

Lai, C. G., G. J. Rix, 2002. "Solution of the Rayleigh eigenproblem in viscoelastic

media." Bulletin of the Seismological Society of America 92(6): 2297-2309. Lucca, Frank, J. 2003, Tight Construction Blasting: Ground Vibration Basics, Monitoring

and Prediction.

Marsden, J.E., D.L. Perkins, and C.C. Krueger. 1995a. Recognition of spawning areas by lake trout: deposition and survival of eggs on small, man-made rock piles. Journal of Great Lakes Research 21 (S1): 330-336.

McDonald, D.G., 1973, The effects of seismic explosions on fish in Sturgeon Lake,

Alberta, Alberta Dept. of Lands & Forests.

Prosser, D. W., 1980, The effects of industrial seismograph operations on water wells, Alberta Environment.

Schneider, L.C., 2001, A survey of blasting vibration regulations, Fragblast, 5, 133-156.

Schmitt, D.R., J. Mwenifumbo, K. A. Pflug, and I.L. Meglis, 2003, Geophysical logging

for elastic properties in hard rock: a tutorial, , in: Hardrock Seismic Exploration (edited by D.W. Eaton, B. Milkereit, and M.H. Salisbury), SEG Geophysical Developments 10, 20-41,

Scott, W.B. and E.J. Crossman. 1973. Freshwater fishes of Canada. Bulletin of the Fisheries Research Board of Canada 184.

Page 115: University of Alberta Library Release Form

98

Stein, S., and M. Wysession, 2003. An introduction to seismology, earthquakes, and earth

structure, Blackwell Pubishing, London, pp. 498.

Telford, W. M., L. P. Geldart, and R. E. Sheriff. 1990. Applied Geophysics. Second Edition. Cambridge University Press. New York.

Welz, M. and D.R. Schmitt. 2005, Effects of Explosives on Incubating Lake Trout Eggs

in the Canadian Arctic. (Blasting Effects Study), DDMI Internal Report, Geophysical Component. 40 pages.

Woods, R.S., and L.P. Jedele, 1985, Energy-Attenuation Relationships from

Construction Vibrations, in Vibration Problems in Geotechnical Engineering (G. Gazetas and E.T. Selig, Eds), Special Technical Publication, ASCE, New York, pp. 187 – 202.

Wright, D.G. and G.E. Hopky. 1998. Guidelines for the use of explosives in or near

Canadian Fisheries waters. Canadian Technical Report of Fisheries and Aquatic Sciences 2107.

Zywicki, D. J. and G. L. Rix 2005. "Mitigation of near-field effects for seismic surface

wave velocity estimation with cylindrical beamformers." Journal of Geotechnical and Geoenvironmental Engineering 131(8): 970-977.

Page 116: University of Alberta Library Release Form

99

Appendices

Appendix 1: Response curves for the Underwater sensor

Response curve for the GS-30CT geophone element used in the 4 Component sensor.

Note that 1 in/sec = 25.4 mm/s.

Page 117: University of Alberta Library Release Form

100

Response curve for the MP25-1200 Hydrophone element used in the 4 Component

sensor. Note that 1 bar = 100 kPa.

Page 118: University of Alberta Library Release Form

101

Appendix 2: Blast Mate blasting record

Typical blasting record as produced by the Instantel MiniBlastmate

Page 119: University of Alberta Library Release Form

102

Appendix 3:

Technical Specifications for the Instantel Minimate Plus Series III.

Page 120: University of Alberta Library Release Form

103

Appendix 4: Blasting Data

DDMI blast data; September 17, 2003 to July 27 2004.

Date Blast Pattern X Y Z Blast Weight (kg) 2004-07-26 370-064 36830.97 53210.99 379.6985 800 2004-07-24 410-011 36998.43 53220.88 415.5052 430 2004-07-22 370-060 36702.85 53220.42 379.5372 750 2004-07-19 410-010 37036.04 53168.81 417.5915 480 2004-07-17 410-009 37049.24 53114 416.5005 480 2004-07-12 390-100 36959.04 52700.47 401.9421 700 2004-07-09 410-005/410-006 37087.8 53091.45 417.7396 800 2004-07-04 400-041 37088.96 53002.46 412.0308 520 2004-06-25 370-057 36385.16 52873.6 380.1243 1200 2004-06-24 405-022b 36947.84 52682.3 405.2773 400 2004-06-22 360-020 36738.47 52806.28 369.72 1500 2004-06-19 380-071 36836.36 52676.96 390.0285 1500 2004-06-14 380-072 36755.22 53277.76 390.5985 520 2004-06-11 380-070 36743.36 53250.25 390.4161 560 2004-06-09 340-027 36537.97 52794.83 351.0323 1100 2004-06-07 380-069 36672.8 52600.5 390.2359 1680 2004-06-06 370-049 36753.65 53287.69 390.9025 560 2004-06-03 370-051 36682.1 52591.05 390.4147 375 2004-05-31 A380-067 36370.19 52761.7 389.9503 1500 2004-05-28 A370-052 36800.95 52783.57 380.184 1500 2004-05-23 A380-064 36387.57 52798.7 389.2335 1000 2004-05-17 A390-096 36439.94 52596.71 407.2856 840

2004-05-14 A370-046/A370-

047 36796.49 52869.12 377.8843 1080 2004-05-10 A380-062 36527.94 52623.16 390.6916 1230 2004-05-08 A380-061 36496.16 52690.46 389.8793 1160 2004-05-06 A370-050 36400.76 52696.91 390.0905 375 2004-05-03 A380-060 36599.36 52646.07 390.3761 2185 2004-05-01 A370-044 36917.49 52878.81 380.3263 1700 2004-04-28 A390-095 36735.14 53295.09 401.1205 660 2004-04-27 A390-094 36853.21 53286.52 405.3811 900 2004-04-25 A340-011 36501.84 52870.94 351.342 2000 2004-04-23 A390-089 36387.74 52642.21 408.1594 1200 2004-04-22 A390-075 36869.6 53288.33 405.5253 150 2004-04-19 A380-059 36615.07 52688.04 390.0126 1010 2004-04-18 A370-041 36879.47 52986.89 380.1734 900 2004-04-13 A390-090 36434.94 52681.49 395.1259 900 2004-04-12 A380-058 36689.11 52659.57 390.1132 1030 2004-04-10 A370-040 36867.97 53067.3 380.0752 520 2004-04-07 A380-057 36696.24 52712.1 390.0146 970 2004-04-04 A395-034 36424.97 52667.6 408.1054 690 2004-03-29 A380-056 36751.52 52726.79 389.9745 2000

Page 121: University of Alberta Library Release Form

104

2004-03-25 A380-055 36783.78 52795.02 387.051 1500 2004-03-22 A390-088 36486.73 52645.02 405.7261 950 2004-03-19 A390-087 36559.93 52663.96 402.2967 765 2004-03-14 A390-086 36684.18 52717.8 397.4487 1040 2004-03-08 A380-053 36948.03 53032.78 390.7004 2500 2004-03-04 A390-085 36565.68 52588.7 400.2385 1880 2004-02-01 A380-052 36856.66 52762.05 390.8498 2000 2004-02-29 A340-004 36600.08 53047.81 351.2579 1500 2004-03-27 A390-083 36627.58 52602.31 400.196 960 2004-04-23 A340-001 36534.18 53237.2 370.2163 71 2004-05-21 A340-003 36639.94 53120.68 351.2147 600 2004-06-20 A350-023 36530.15 52787.62 360.3096 1085 2004-06-18 A340-002 36552.64 53078.2 350.6155 920 2004-02-15 390-082 36691.39 52619.27 400.2983 900 2004-02-12 380-050 36927 52859 391 660 2004-02-09 350-022 36459 52851 360 1500 2004-02-07 390-081 36744 52651 401 2000 2004-02-01 390-080 36808 52662 402 1400 2004-01-17 400-038 36575 52560 406 300 2004-01-07 390-072 36925 53219 409 580 2004-01-05 360-017 36462 52844 370 520 2004-01-01 350-014 36566 52982 361 3000 2003-12-28 350-011 36729 52959 361 1000 2003-12-26 370-038 36429 52810 387 850 2003-12-23 370-037 36422 52897 382 1300 2003-12-15 350-010 36613 52992 361 2000 2003-12-11 370-036 36802 53215 380 1800 2003-12-07 350-007 36712 53059 360 500 2003-11-30 350-006 36690 52985 361 1010 2003-11-29 370-034 36383 53073 380 490 2003-11-26 350-005 36638 53060 360 1080 2003-11-23 370-033 36488 53213 380 490 2003-11-20 350-004 36582 53118 360 1090 2003-11-19 370-035 36608 53219 380 490 2003-11-16 370-032 36510 53195 380 500 2003-11-14 350-003 36716 53117 360 800 2003-11-12 380-044 36882 53009 390 1000 2003-11-09 350-001 36628 53125 360 1500 2003-11-08 390-066 37000 53072 405 700 2003-11-06 380-043 36551 53246 390 500 2003-11-03 370-031 36402 53033 380 500 2003-11-01 370-029 PS 36539 53256 390 510 2003-10-27 360-012 36798 53061 370 1100 2003-10-25 360-013 36572 53115 371 1100 2003-10-22 360-011 36690 53119 370 1000 2003-10-19 380-042 36355 53008 390 480 2003-10-17 380-041 36488 53184 390 560 2003-10-11 370-028 36674 53167 380 510

Page 122: University of Alberta Library Release Form

105

2003-10-09 360-008 36738 53022 370 1700 2003-10-02 380-040 36355 53008 390 400 2003-09-27 360-005 36511 52928 370 1540 2003-09-25 360-006 36706 52959 370 1110 2003-09-22 390-063 36496 53254 395 280 2003-09-20 360-004 36634 53037 370 850 2003-09-17 385-039 36529 53179 390 1000 2003-09-13 370-025 36372 53092 390 1500 2003-09-07 380-038 36870 53087 390 565 2003-09-06 370-024 36760 52915 380 1875 2003-09-03 370-023 36699 52848 380 1000

U of A blast data; September 17, 2003 to July 27 2004.

South Dike (8902) Ref. Island (8903) Tern Island (8904) East Dike (8905) Date Dist. SD PPV Dist. SD PPV Dist. SD PPV Dist. SD PPV

26-Jul-04 871.2 30.8 1.98 2804.7 99.2 956.1 33.8 711.0 25.1 5.83 24-Jul-04 844.7 40.7 2637.2 127.2 865.1 41.7 554.8 26.8 22-Jul-04 927.0 33.8 1.89 2930.8 107.0 0.635 1050.2 38.3 837.1 30.6 3.65 19-Jul-04 789.2 36.0 1.77 2607.2 119.0 801.1 36.6 501.8 22.9 6.52 17-Jul-04 733.6 33.5 4.49 2602.9 118.8 747.5 34.1 475.0 21.7 6.51 12-Jul-04 347.2 13.1 2790.0 105.5 537.0 20.3 638.4 24.1 11.5 9-Jul-04 709.5 25.1 2.6 2568.9 90.8 708.1 25.0 432.9 15.3 9.15 4-Jul-04 620.5 27.2 2.14 2585.4 113.4 632.6 27.7 425.3 18.6 6.14

25-Jun-04 865.6 25.0 6.33 3300.7 95.3 1134.6 32.8 1138.2 32.9 3.25 24-Jun-04 335.3 16.8 2806.3 140.3 540.3 27.0 657.2 32.9 22-Jun-04 556.9 14.4 2972.4 76.7 782.6 20.2 804.4 20.8 19-Jun-04 394.1 10.2 12.9 2914.1 75.2 643.9 16.6 758.0 19.6 8.7 14-Jun-04 958.9 42.1 9.05 2872.5 126.0 1055.8 46.3 802.9 35.2 11-Jun-04 937.7 39.6 8.5 2887.2 122.0 1043.3 44.1 805.7 34.0 9-Jun-04 697.4 21.0 7.39 3169.8 95.6 965.7 29.1 1002.5 30.2 3.21 7-Jun-04 477.6 11.7 9.87 3093.2 75.5 786.3 19.2 938.4 22.9 4.72 6-Jun-04 968.8 40.9 4.47 2873.2 121.4 1064.3 45.0 807.7 34.1 7.83 3-Jun-04 465.0 24.0 10.9 3087.3 159.4 775.7 40.1 934.4 48.2 7.98

31-May-04 820.3 21.2 6.16 3340.2 86.2 1115.4 28.8 1171.9 30.3 28-May-04 499.8 12.9 2917.9 75.3 715.9 18.5 7.98 750.6 19.4 23-May-04 823.0 26.0 9.42 3314.6 104.8 1108.5 35.1 3.53 1147.4 36.3 17-May-04 691.3 23.9 5.81 3316.4 114.4 1016.4 35.1 2.38 1153.0 39.8 1.95 14-May-04 573.3 17.4 2900.7 88.3 759.7 23.1 5.79 733.1 22.3 9.57 10-May-04 618.4 17.6 4.95 3224.6 91.9 932.8 26.6 2.5 1061.6 30.3 2.11 8-May-04 675.7 19.8 4.47 3236.4 95.0 976.4 28.7 2.79 1068.8 31.4 6-May-04 764.4 39.5 3326.7 171.8 1071.0 55.3 1158.2 59.8 3-May-04 563.7 12.1 10.5 3149.6 67.4 866.3 18.5 986.8 21.1 5.77 1-May-04 529.1 12.8 13.4 2780.9 67.4 664.2 16.1 15.6 612.8 14.9 12.6 28-Apr-04 982.2 38.2 2890.9 112.5 1081.9 42.1 827.4 32.2 27-Apr-04 936.8 31.2 12.2 2774.1 92.5 1000.9 33.4 714.0 23.8 12.2 25-Apr-04 772.5 17.3 3187.6 71.3 1026.1 22.9 10.3 1024.2 22.9 7.31

Page 123: University of Alberta Library Release Form

106

23-Apr-04 755.5 21.8 3353.7 96.8 1074.1 31.0 7.48 1186.7 34.3 22-Apr-04 934.4 76.3 6.61 2757.6 225.2 992.7 81.1 5.31 699.6 57.1 6.81 19-Apr-04 571.2 18.0 7.03 3122.6 98.3 860.0 27.1 5.23 957.1 30.1 4.14 18-Apr-04 643.5 21.5 8.26 2793.7 93.1 761.8 25.4 635.9 21.2 7.67 13-Apr-04 726.8 24.2 5.41 3297.7 109.9 1034.4 34.5 1129.8 37.7 2.74 12-Apr-04 493.8 15.4 3059.8 95.3 781.4 24.3 898.8 28.0 10-Apr-04 723.2 31.7 7.7 2789.7 122.3 825.3 36.2 8.14 648.8 28.4 9.11 7-Apr-04 519.6 16.7 3038.0 97.5 788.1 25.3 872.6 28.0 4.63 4-Apr-04 730.2 27.8 3311.0 126.0 1041.5 39.7 1143.4 43.5 4.04

29-Mar-04 488.5 10.9 24.9 2980.8 66.7 0.508 740.1 16.5 17.4 815.7 18.2 11.2 25-Mar-04 518.9 13.4 8.16 2931.3 75.7 736.0 19.0 763.3 19.7 4.64 22-Mar-04 664.5 21.6 9.9 3257.8 105.7 976.8 31.7 3.5 1092.2 35.4 3.78 19-Mar-04 606.6 21.9 5.27 3182.2 115.1 0.508 908.3 32.8 9.63 1017.0 36.8 4.44 14-Mar-04 532.3 16.5 7.44 3047.9 94.5 801.1 24.8 4.71 881.9 27.3 5.39

8-Mar-04 667.9 13.4 15.6 2717.3 54.3 744.7 14.9 566.5 11.3 19.2 4-Mar-04 570.2 13.2 3198.7 73.8 890.7 20.5 1040.2 24.0 1-Feb-04 450.1 10.1 2870.0 64.2 656.0 14.7 704.9 15.8

29-Feb-04 832.9 21.5 3057.4 78.9 1022.2 26.4 916.4 23.7 27-Mar-04 518.6 16.7 1.57 3135.6 101.2 831.1 26.8 978.3 31.6 3.95 23-Apr-04 1024.6 121.6 3.3 3096.7 367.5 1187.1 140.9 1005.3 119.3 9.56

21-May-04 870.6 35.5 7.21 3006.6 122.7 1032.6 42.2 882.3 36.0 3.14 20-Jun-04 698.8 21.2 2.88 3179.0 96.5 970.3 29.5 1011.2 30.7 18-Jun-04 885.8 29.2 5.36 3099.3 102.2 1078.4 35.6 965.2 31.8 5.05 15-Feb-04 470.0 15.7 3069.7 102.3 770.8 25.7 913.3 30.4 5.69 12-Feb-04 506.7 19.7 8.32 2776.5 108.1 644.4 25.1 607.9 23.7 7.61 9-Feb-04 793.4 20.5 6.21 3233.6 83.5 1058.3 27.3 1069.0 27.6 3.43 7-Feb-04 443.9 9.9 14.6 3009.9 67.3 725.8 16.2 852.2 19.1 10.9 1-Feb-04 402.5 10.8 19.5 2945.6 78.7 666.5 17.8 12.7 789.8 21.1 12.1

17-Jan-04 551.5 31.8 4.69 3198.7 184.7 0.582 878.4 50.7 10.7 1043.9 60.3 6.13 7-Jan-04 854.5 35.5 6.62 2710.3 112.5 904.1 37.5 5.3 623.0 25.9 8.52 5-Jan-04 786.3 34.5 6.29 3232.1 141.7 1052.6 46.2 3.41 1066.7 46.8 2.95 1-Jan-04 802.7 14.7 0 3102.3 56.6 1016.3 18.6 5.55 949.9 17.3

28-Dec-03 686.1 21.7 7.36 2946.9 93.2 0.696 865.3 27.4 6.18 788.6 24.9 5.55 26-Dec-03 793.7 27.2 3.99 3271.7 112.2 1072.3 36.8 1104.6 37.9 23-Dec-03 849.5 23.6 6.56 3259.8 90.4 1108.6 30.7 1098.8 30.5 7.48 15-Dec-03 780.2 17.4 4.49 3054.3 68.3 980.8 21.9 902.7 20.2 4.79 11-Dec-03 884.2 20.8 12.5 2833.0 66.8 978.1 23.1 739.9 17.4 28.5 7-Dec-03 780.4 34.9 4.6 2945.0 131.7 937.5 41.9 804.7 36.0 5.21

30-Nov-03 729.1 22.9 4.67 2980.0 93.8 912.1 28.7 826.0 26.0 5.18 29-Nov-03 994.1 44.9 0 3267.3 147.6 1219.6 55.1 1132.8 51.2 26-Nov-03 820.3 25.0 6.51 3017.8 91.8 997.4 30.3 878.5 26.7 4.12 23-Nov-03 1030.7 46.6 2.42 3145.1 142.1 1208.0 54.6 1045.2 47.2 2.29 20-Nov-03 899.7 27.3 3.77 3064.2 92.8 1076.6 32.6 938.9 28.4 4.38 19-Nov-03 969.8 43.8 1.99 3025.2 136.7 1118.7 50.5 928.9 42.0 2.98 16-Nov-03 1003.2 44.9 0 3125.3 139.8 1179.6 52.8 1020.3 45.6 14-Nov-03 829.4 29.3 0 2931.8 103.7 971.2 34.3 805.9 28.5 12-Nov-03 663.1 21.0 0 2786.7 88.1 774.4 24.5 632.4 20.0 9-Nov-03 880.1 22.7 2.54 3017.6 77.9 1044.2 27.0 894.1 23.1 4.25

Page 124: University of Alberta Library Release Form

107

8-Nov-03 696.8 26.3 2.5 2658.8 100.5 741.2 28.0 517.1 19.5 10.8 6-Nov-03 1022.3 45.7 2.09 3078.9 137.7 1179.3 52.7 990.3 44.3 2.17 3-Nov-03 952.7 42.6 0 3254.8 145.6 1183.9 52.9 1112.5 49.8 1-Nov-03 1037.1 45.9 4.88 3089.9 136.8 1194.9 52.9 1004.3 44.5 2.79

27-Oct-03 743.0 22.4 3.11 2859.8 86.2 872.2 26.3 718.5 21.7 5.65 25-Oct-03 902.5 27.2 1.82 3074.3 92.7 1082.5 32.6 948.0 28.6 2.1 22-Oct-03 842.9 26.7 2.58 2957.0 93.5 991.8 31.4 4 831.4 26.3 4.15 19-Oct-03 971.0 44.3 2.08 3305.1 150.9 1214.5 55.4 1159.2 52.9 1.86 17-Oct-03 1007.2 42.6 4.42 3148.4 133.0 1190.5 50.3 1039.9 43.9 11-Oct-03 892.2 39.5 1.84 2966.0 131.3 1034.9 45.8 854.0 37.8 2.9 9-Oct-03 734.9 17.8 0 2925.8 71.0 893.9 21.7 777.1 18.8 2-Oct-03 971.0 48.5 0 3305.1 165.3 1214.5 60.7 1.25 1159.2 58.0

27-Sep-03 801.9 20.4 3.87 3166.6 80.7 0.582 1039.9 26.5 4.37 1007.6 25.7 3.1 25-Sep-03 698.1 21.0 0 2969.3 89.1 884.1 26.5 811.0 24.3 22-Sep-03 1059.2 63.3 0 3132.8 187.2 1226.7 73.3 1045.6 62.5 20-Sep-03 803.1 27.5 3.44 3025.5 103.8 987.1 33.9 881.2 30.2 17-Sep-03 978.9 31.0 0 3108.3 98.3 1154.6 36.5 998.6 31.6 13-Sep-03 1015.0 26.2 0 3275.4 84.6 1238.3 32.0 1144.7 29.6 7-Sep-03 740.9 31.2 3.27 2784.2 117.1 0.582 837.6 35.2 5.43 648.2 27.3 5.71 6-Sep-03 631.3 14.6 7.51 2925.7 67.6 0.907 814.6 18.8 10.3 761.3 17.6 8.64 3-Sep-03 613.6 19.4 5.42 3000.4 94.9 0.582 835.5 26.4 6.03 832.7 26.3 4.01